Development and validation of a nomogram to predict the risk of renal replacement therapy among acute kidney injury patients in intensive care unit

Abstract

Abstract Background There are no universally accepted indications to initiate renal replacement therapy (RRT) among patients with acute kidney injury (AKI). This study aimed to develop a nomogram to predict the risk of RRT among AKI patients in intensive care unit (ICU). Methods In this retrospective cohort study, we extracted AKI patients from Medical Information Mart for Intensive Care III (MIMIC-III) database. Patients were randomly divided into a training cohort (70%) and a validation cohort (30%). Multivariable logistic regression based on Akaike information criterion was used to establish the nomogram. The discrimination and calibration of the nomogram were evaluated by Harrell’s concordance index (C-index) and Hosmer–Lemeshow (HL) test. Decision curve analysis (DCA) was performed to evaluate clinical application. Results A total of 7413 critically ill patients with AKI were finally enrolled. 514 (6.9%) patients received RRT after ICU admission. 5194 (70%) patients were in the training cohort and 2219 (30%) patients were in the validation cohort. Nine variables, namely, age, hemoglobin, creatinine, blood urea nitrogen and lactate at AKI detection, comorbidity of congestive heart failure, AKI stage, and vasopressor use were included in the nomogram. The predictive model demonstrated satisfying discrimination and calibration with C-index of 0.938 (95% CI, 0.927–0.949; HL test, P = 0.430) in training set and 0.935 (95% CI, 0.919–0.951; HL test, P = 0.392) in validation set. DCA showed a positive net benefit of our nomogram. Conclusion The nomogram developed in this study was highly accurate for RRT prediction with potential application value.