Risk Factors and Machine Learning Model for Aminoglycines Related Acute Kidney Injury

Aminoglycoside Toxicity, Acute Kidney Injury

Aminoglycosides, Acute kidney injury, Pharmacoepidemiology, Miss diagnosis, Risk factors, Logic regression

Drug-induced acute kidney injury (D-AKI) can occur after treatment with aminoglycosides. Predicting the risk of D-AKI is important for a tailored prevention and palliation strategy. There are currently no studies to construct a model for predicting the risk of D-AKI associated with aminoglycosides. Therefore, the study aimed to develop a model to predict the risk of D-AKI that could be used in clinical practice. Clinical data of inpatients treated with aminoglycosides at the First Affiliated Hospital of Shandong First Medical University from January 2018 to December 2020, were collected. The primary endpoint was D-AKI, defined according to the 2012 Global Outcomes for Kidney Disease Improvement (KDIGO). Patient clinical information, including demographic information, admission and discharge information, disease history, medication information, and laboratory tests, was obtained through an in-hospital electronic medical record system. Independent risk factors associated with D-AKI will be screened by univariate and multifactorial analyses. Covariates with significant differences (P \< 0.05) were included in logistic regression models. The models were evaluated by the area under the curve (AUC) of the receiver operating characteristic curve (ROC) obtained by ten-fold cross-validation. Future studies are needed to test the application of this model in clinical practice to determine whether D-AKI in this setting can be predicted and mitigated.

China