జర్నల్ ఆఫ్ క్లినికల్ ట్రయల్స్
అందరికి ప్రవేశం

నైరూప్య

Development and Assessment of an Efficient Predictive Nomogram for Bloodstream Infection in Gallstone Patients

Xiuqing Shen, Changsheng Wu, Rongxin Zhong, Falin Chen, Sijie Wang, Pengju Cao*, Shaoting Chen*

Objective: The objective of this study was to create and authenticate a user-friendly nomogram for assessing the risk of Bloodstream Infections (BSI) in patients with gallstones.

Materials and methods: Data on 261 patients without BSI and 96 patients with BSI who had gallstones were gathered between January 2019 and April 2023. The ten variables—age, sex, White Blood Cell (WBC), Neutrophil (NE%), calcitonin (PCT), γ-Glutamyltransferase (GGT), Total Bilirubin (TBIL), Direct Bilirubin (DBIL), Alkaline Phosphatase (ALP), and glutamic oxaloacetic transaminase (AST)—were first screened using the single-factor regression analysis. Multiple regression analyses were used to optimize the risk model's predictive features. The prediction model's discrimination, calibration, and clinical application performance were evaluated using the consistency index (c-index), calibration curve, and Clinical Decision Curve (DCA).

Results: Nine variables show no multicollinearity, with the exception of gender, according to collinearity analysis. Multiple logistic regression analysis was included to select age, PCT, and AST predictors. Among them, age [OR=0.562 (0.366–0.939), P=0.028] is a protective factor for gallstone patients with blood flow infection (P<0.05), whereas PCT [OR=2.115 (1.244–3.597), P=0.006], and AST [OR=3.469 (1.942–6.198), P=0.000] are risk factors for concurrent BSI in gallstone patients with gallstones (P<0.05). On the basis of the anticipated variables, a prediction model was created. With a bootstrap validation of 0.7 and a C-index of 0.71, the prediction model's validation shows sufficient discrimination. The calibration curve showed how well the model can be corrected and how accurate its predictions are. The DCA results demonstrated that the predictive model was clinically effective when the threshold probability was between 17%–77%, as the net return was more significant than the two extreme models.

Conclusion: As an additional technique for predicting individuals with gallstones complicated by blood flow infection, the nomogram has some predictive power.