Constructing a predictive model for urinary tract infections within one year after kidney transplantation from donation after brain death based on machine learning
10.3760/cma.j.cn114452-20240607-00304
- VernacularTitle:基于机器学习构建供者脑死亡后器官捐献肾移植术后1年内泌尿系感染的预测模型
- Author:
Yang CHU
1
;
Feng REN
1
Author Information
1. 大连医科大学附属第二医院检验科PCR实验室，大连 116023
- Publication Type:Journal Article
- Keywords:
Kidney transplantation;
Urinary tract infection;
Machine learning;
Model prediction
- From:
Chinese Journal of Laboratory Medicine
2025;48(3):331-336
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To develop a predictive model for urinary tract infection (UTI) within one year after kidney transplantation (KT) from donation after brain death (DBD).Methods:A retrospective case-control study was conducted on 77 DBD kidney transplant recipients at the Second Affiliated Hospital of Dalian Medical University from June 1, 2021, to January 31, 2023. Patients were divided into a UTI group ( n=31) and a non-UTI group ( n=46) based on the occurrence of UTI within one year post-transplant. In the UTI group, males and females accounted for 58% and 42%, respectively, with a mean age of (42.5±11.6) years. In the non UTI group, males and females accounted for 72% and 28%, respectively, with a mean age of (44.0±10.2) years. No significant differences were observed in gender ( P=0.213) or age ( P=0.548) between the two groups. Basic information of the KT recipients, as well as laboratory indicators such as renal function and procalcitonin, were collected. Logistic regression was used for univariate and multivariate analyses to identify risk factors. Machine learning (ML) was employed to construct a prediction model and a nomogram. For the final prediction model, a random 20% sample was drawn from the total sample as the test set, with the remaining samples used as the training set for 10-fold cross-validation. Results:Univariate regression and correlation analyses revealed that pre-discharge uric acid ( OR=1.009, 95% CI 1.003-1.014, P=0.004), urea ( OR=1.283, 95% CI 1.096-1.502, P=0.002), and procalcitonin ( OR=2.797, 95% CI 1.006-7.772, P=0.049) were associated with the occurrence of UTI. Based on these findings, we constructed a prediction model for UTI following DBD kidney transplantation. The logistic regression model with the largest area under the receiver operating characteristic curve (AUC) was selected as the final prediction model, which had an AUC of 0.841 (95% CI 0.619-1.000), with a sensitivity and specificity of 0.857 and 0.889, respectively. Conclusions:In summary, urea, uric acid, and procalcitonin are risk factors for the occurrence of UTI within one year after DBD kidney transplantation. The logistic regression model constructed based on these indicators has clinical application value in predicting the risk of UTI within one year after DBD kidney transplantation.
