Brian Robert Herts MD, Presenter: Researcher, Siemens AG

Mark Early Baker MD, Abstract Co-Author: Medical Advisory Board, E-Z-EM, Inc Speakers Bureau, E-Z-EM, Inc

Nancy A. Obuchowski PhD, Abstract Co-Author: Consultant, Siemens AG

Emilio Poggio MD, Abstract Co-Author: Nothing to Disclose

Erika Schneider PhD, Abstract Co-Author: Nothing to Disclose

To determine a method for predicting risk of contrast-induced nephropathy (CIN) after contrast-enhanced CT, based on creatinine, age, gender, race or the estimated GFR (eGFR) which is a combination of these variables.

Of 5187 outpatients who underwent contrast-enhanced CT (CECT), 2739 had serum creatinine (sCr) within 6 months before the exam. Along with age, gender, and race, sCr was used to calculate pre-CT eGFR using the 4-variable MDRD. 546 patients also had sCr within 2 weeks following the CT; the minimum post-CT eGFR was calculated. Pre-CT sCr, age, race, gender and eGFR were tested using multiple-variable logistic regression models generating ROC curves to predict reduced renal function using criteria of eGFR of 0.5 mg/dl or >25%) were also correlated with reduced eGFR after CT.

There were 69(12.6%) patients with eGFR <75 after CT. Race and gender were not relevant in the model. Pre-CT eGFR alone (model 1, ROC area 0.839, p<0.001) and the combination of age and sCr (model 2, ROC area 0.854, p<0.002) were significant predictors of eGFR <75 after CT. The probability of CIN for any patient is exp(y)/(1+exp(y)), where y=3.68-(0.057 x eGFR)[Model 1] or y= -7.865+(0.033 x age)+(3.617 x creat) [Model 2]. For example, the average probability of eGFR<75 following CT was 19% for eGFR<90 and 36% for eGFR<75 pre-CT. Alternatively, the average probability of post-CT eGFR<75 for a 70 year old with sCr 1.0 mg/dl was 40%. The number of patients (n=17,3.1%) with eGFR <60 was too small for statistical modeling. There was poor correlation of the traditional CIN indicators with reduced eGFR.

Using conservative criteria for CIN of eGFR<75, probability of CIN after CT can be predicted by a combination of age and creatinine or eGFR using the 4-variable MDRD. Criteria need to be established for what is an acceptable risk in order to manage patients undergoing CECT. Furthermore, standard criteria for CIN are poor predictors of renal insufficiency following CT.

Accurate knowledge of who is at risk for CIN will help determine who gets renal protective strategies before contrast infusions.

Herts, B, Baker, M, Obuchowski, N, Poggio, E, Schneider, E, Contrast-induced Nephropathy Following CT: Creating a Model to Predict Probability Based on Creatinine, Age, and Estimated GFR.  Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL. http://archive.rsna.org/2007/5013550.html