Clinical question: Do newer prediction models pooled in a systematic review predict contrast-associated acute kidney injury (CA-AKI)?
Background: This systematic review updates a previous evaluation (Silver SA, et al. Risk prediction models for contrast-induced nephropathy: systematic review. BMJ. 2015;351:h4395. doi:10.1136/bmj.h4395). Since then, the number of CA-AKI models has quadrupled. Silver’s systematic review noted the best performing models included age, diabetes, chronic kidney disease, heart failure, and hypotension, with an area under the curve of 0.82 (95% CI, 0.81 to 0.83), and were only relevant to contrast given for coronary angiography.
Study design: Systematic review and meta-analysis
Synopsis: Of 64 models for the primary outcome of CA-AKI (as defined in each included study), 45 studies were included; only nine models (14.1 %) had low risk of bias, enabling a summary receiver operating characteristic curve analysis with a pooled C statistic, area under the curve of 0.83 (95% CI, 0.82 to 0.84); sensitivity, 0.74 (CI, 0.7 to 0.78); specificity, 0.78 (95% CI, 0.75 to 0.82). The five most used variables were baseline renal function, age, coronary artery disease, cardiac function, and past medical history (cardiac, diabetes mellitus, peripheral vascular disease, and stroke).
Bottom line: We do not have a current preferred or clinically useful model to predict CA-AKI. Limitations for models include the retrospective nature of studies and bias induced by the exclusion of patients with more advanced CKD who never underwent contrast.
Citation: Feng Y, et al. Predicting contrast-associated acute kidney injury. JAMA Netw Open. 2025;8(3):e250107. doi:10.1001/jamanetworkopen.2025.0107.
Dr. Mohan is a hospitalist in the department of hospital medicine at the Cleveland Clinic and assistant professor of medicine at Cleveland Clinic Lerner College of Medicine of Case Western Reserve University, both in Cleveland.