A novel patient-centered approach for defining contrast thresholds to reduce contrast-induced kidney injury from percutaneous coronary intervention.

Background:
Contrast-Induced Acute Kidney Injury (CI-AKI) is the most common non-cardiac complication of PCI and the best method for reducing CI-AKI is to minimize contrast. The most common method for estimating a ‘safe’ contrast dose is to restrict the volume to less than 3 × GFR. We sought to leverage the validated AKI NCDR risk model to back calculate the contrast needed to reduce AKI by 10% among patients with above average risk (>7%) and compared the impact of this threshold on AKI rates, stratified by contrast volumes above and below 3 × GFR.

Methods:
Using NCDR CathPCI version 4 data from June 2009-June 2012 in 391,819 patients with creatinine levels before and after PCI, we compared the contrast limits and AKI rates among patients adhering to the NCDR model, stratified by 3 × GFR. AKI was defined an increase of 0.3mg/dl, in accordance with the AKIN criteria. We further stratified the cohort into moderate (7-10%), high (10-15%) and very high (>15%) AKI risk.

Results:
Figure 1 shows the difference in contrast volume calculated from our method and 3 × GFR, along with the rates of AKI when model thresholds were exceeded. While the novel approach generally allowed less contrast than 3 × GFR, the rates of AKI, both when the 3 × GFR rate was and was not exceeded, were significantly lower when the NCDR risk model-based contrast limits were not exceeded. The difference was strongest in patients at the highest risk.

Conclusions:
Using the NCDR AKI risk model to estimate safe contrast limits resulted in a lower rate of CI AKI than contrast limits estimated by 3 × GFR. This strategy should be tested to improve the safety of PCI.