Contrast-induced acute kidney injury (CIAKI) is a frequent complication after infusion of contrast media in patients undergoing percutaneous coronary intervention. A wide range of CIAKI rates occurs after intervention between 3% and 30%, depending on the definition. The aim of this study was to identify which methodology was more effective at recognizing patients at high risk for in-hospital and out-of-hospital adverse events.

Serum creatinine increases, after contrast agent infusion, were evaluated in 755 consecutive and unselected patients. Incidences of CIAKI diagnosed by 2 common definitions varied from 6.9% (creatinine increase of ≥0.5 mg/dL, CIAKI-0.5) to 15.9% (creatinine increase of ≥25%, CIAKI-25%). Significant differences appeared between the 2 definitions of sensitivity to predict renal failure according to receiver operating characteristic curve analysis (98% for CIAKI-0.5 and 62% for CIAKI-25%), using a cutoff value of postprocedural glomerular filtration rate of 60 mL/min. Both definitions of CIAKI were related to composite adverse events, but CIAKI-0.5 showed a stronger predicting value (odds ratio 2.875 vs 1.802, P = .036). In multivariate linear regression, only CIAKI-0.5 was a predictive variable of death (odds ratio 3.174, 95% CI 1.368-7.361).

An increase in serum creatinine of ≥0.5 mg/dL is more sensitive because it recognizes more selectively those patients with a higher risk of mortality and morbidity. Serum creatinine increases of ≥25% overestimate CIAKI by including many patients without postprocedural relevant deterioration of renal function and affected by a lower risk of adverse events at follow-up.