What question did this study set out to answer?

The central aim is to develop a prediction model for assessing acute kidney injury risk in in-hospital cardiac arrest patients during ICU stays.

May 3, 2026SHILAP Revista de lepidopterología0 citationsOpen Access

Development of a Clinical Prediction Model for Acute Kidney Injury Among In-Hospital Cardiac Arrest Patients During Intensive Care Unit Hospitalization

HAHalidan Abudu ZLZiquan Liu YNYanxiang Niu

Key Points

The central aim is to develop a prediction model for assessing acute kidney injury risk in in-hospital cardiac arrest patients during ICU stays.
Used the MIMIC-IV database to create a nomogram for AKI prediction based on clinical characteristics.
Conducted variable selection using LASSO regression, followed by univariate and multivariate logistic regression analyses.
Evaluated model performance through sensitivity, specificity, Youden index, DCA, CIC, and ROC curve analysis.
Included 1427 cardiac arrest patients: training cohort (n=999) and validation cohort (n=428).
Identified five independent predictors for AKI: weight, SpO2, sodium, SOFA score, and OASIS score with respective aOR values and 95% CIs.
Model AUCs were 0.920 (training) and 0.875 (validation), with a specificity of 0.837 and sensitivity of 0.781.

Abstract

Background: Acute kidney injury (AKI) is a significant cause of mortality among post-cardiac arrest patients. However, clinical prediction models for assessing AKI risk for in-hospital cardiac arrest (IHCA) patients remain limited. Thus, this retrospective study aimed to develop a nomogram that uses readily available clinical characteristics to predict the likelihood of AKI in this group of patients during intensive care unit (ICU) hospitalization. Methods: This study constructed a nomogram based on the Medical Information Mart for Intensive Care IV (MIMIC-IV) database and conducted variable selection through Least Absolute Shrinkage and Selection Operator (LASSO) regression, followed by univariate and multivariate logistic regression analyses on the selected variables. Model performance was evaluated by calculating sensitivity, specificity, and the Youden index, and by performing decision curve analysis (DCA), clinical impact curve (CIC), and receiver operating characteristic (ROC) curve analysis. Results: This study included 1427 cardiac arrest (CA) patients, who were randomly allocated into a training cohort (n = 999) and a validation cohort (n = 428). We identified five independent predictors for post-cardiac arrest AKI: weight (adjusted odds ratio (aOR): 1.016, 95% confidence interval (CI): 1.009–1.024), peripheral capillary oxygen saturation (SpO2) (aOR: 1.044, 95% CI: 1.026–1.063), sodium (aOR: 0.947, 95% CI: 0.919–0.975), Sequential Organ Failure Assessment (SOFA) score (aOR: 1.134, 95% CI: 1.083–1.190), and Oxford Acute Severity of Illness Score (OASIS) score (aOR: 1.080, 95% CI: 1.059–1.103). The model demonstrated strong performance, with area under the curve (AUC) values of 0.920 and 0.875 in the training and validation cohorts, respectively. Upon validation, the specificity, sensitivity, and Youden index for the model were 0.837, 0.781, and 0.618, respectively. The calibration curve indicated good agreement between predictions and observations. The DCA and CIC confirmed the clinical utility of the model. Conclusion: The developed prediction model exhibits high predictive performance for predicting AKI in IHCA patients.

AI에게 질문

Bookmark

View Full Paper