Development and validation of a predictive model for post-percutaneous nephrolithotomy urinary sepsis: a multicenter retrospective study.

Abstract

Background: The objective of this retrospective, multicenter study was to analyze the factors associated with the development of urogenital sepsis after percutaneous nephrolithotomy (PCNL) and to establish a nomogram prediction model of urogenital sepsis after PCNL.

Methods: A total of 2066 postoperative PCNL patients were included from three medical institutions: Zunyi Medical University Hospital, Beijing Jishuitan Hospital Guizhou Hospital, and Fenggang County People's Hospital. Clinical data of 1623 patients from the Department of Urology of Zunyi Medical University Hospital were randomized into a training cohort (Zunyi training cohort, N.=1139) and an internal validation cohort (Zunyi internal validation cohort, N.=484) using computer generated random numbers in a 7:3 ratio. Univariate and multivariate logistic regression analyses were performed on the compliance training cohort to identify risk factors for urogenital sepsis after PCNL and to develop a column line graph prediction model based on these risk factors. Finally, Zunyi internal validation cohort and two external validation cohorts (Guiyang external cohort, N.=306; Fenggang external cohort, N.=137) were used to validate the prognostic accuracy of the nomogram prediction model. R4.2.2 statistical software was used for all statistical data analyses.

Results: Multifactorial logistic regression analysis of the Zuiyi training cohort (N.=1139) identified five independent risk factors associated with urogenital sepsis after PCNL, including urine culture positivity (odds ratio [OR]=5.29, P<0.001), urine nitrite positivity (OR=5.97, P<0.001), operation time ≥60 min (OR=4.4, P=0.0037), residual stone (OR=5.18, P<0.001), and size ≥30 mm (OR=3.22, P=0.0086). Nomogram were constructed based on these independent risk factors. The area under the curve (AUC) of the nomogram model was 0.907 in the in-progress sample and 0.948 after internal validation. The AUC of the model was 0.855 and 0.804 after external validation of the Guiyang external validation cohort and the Fenggang validation cohort, respectively, indicating good discrimination ability. The calibration curves of the nomogram showed good agreement, and the decision curve analysis demonstrated high clinical utility.

Conclusions: Based on the clinical independent risk factors such as positive urine culture, positive urine nitrite, operation time ≥60min, stone residue, stone size ≥30mm, nomogram prediction model of urogenital sepsis after PCNL was established, which can provide reference for urologists to develop preoperative evaluation and treatment strategies for patients with percutaneous nephrolithotomy.