Construction and evaluation of a risk prediction model for pulmonary infection‐associated acute kidney injury in intensive care units

Acute kidney injury (AKI) is one of the common complications of pulmonary infections. However, nomograms predicting the risk of early‐onset AKI in patients with pulmonary infections have not been comprehensively researched. In this study, 3278 patients with pulmonary infection were extracted from the Medical Information Mart for Intensive Care III (MIMIC‐III) database. These patients were randomly divided into training and validation cohorts, with the training cohort used for model building and the validation cohort used for validation. Independent risk factors for patients with pulmonary infection were determined using the least absolute shrinkage and selection operator (LASSO) method and forward stepwise logistic regression, which revealed that 11 independent risk factors for AKI in patients with pulmonary infections were congestive heart failure (CHF), hypertension, diabetes, transcutaneous oxygen saturation (SpO2), 24‐h urine output, white blood cells (WBC), serum creatinine (Scr), prothrombin time (PT), potential of hydrogen (PH), vasopressor use, and mechanical ventilation (MV) use. The nomogram was then constructed and validated. The area under the receiver operating characteristic curve (AUC) values of the nomogram were 0.770 (95% CI = 0.789–0.807) in the training cohort and 0.724 (95% CI = 0.754–0.784) in the validation cohort. High AUC values indicated the good discriminative ability of the nomogram, while the calibration curves and Hosmer–Lemeshow test results indicated that the nomogram was well‐calibrated. Improvements in net reclassification index (NRI) and integrated discrimination improvement (IDI) values indicate that our nomogram was superior to the Simplified Acute Physiology Score (SAPS) II scoring system, and the decision–curve analysis (DCA) curves indicate that the nomogram has good clinical application. We established a risk‐prediction model for AKI in patients with pulmonary infection, which has good discriminative power and is superior to the SAPS II scoring system. This model can provide clinical reference information for patients with this type of disease in the intensive care unit.