Prediction of presence of kidney disease in a general patient population undergoing intravenous iodinated contrast enhanced computed tomography

To assess which risk factors can be used to reduce superfluous estimated glomerular filtration rate (eGFR) measurements before intravenous contrast medium administration. In consecutive patients, all decreased eGFR risk factors were assessed: diabetes mellitus (DM), history of urologic/nephrologic disease (HUND), nephrotoxic medication, cardiovascular disease, hypertension, age > 60 years, anaemia, malignancy and multiple myeloma/M. Waldenstroem. We studied four models: (1) all risk factors, (2) DM, HUND, hypertension, age > 60 years; (3) DM, HUND, cardiovascular disease, hypertension; (4) DM, HUND, age > 75 years and congestive heart failure. For each model, association with eGFR < 60 ml/min/1.73 m2 or eGFR < 45 ml/min/1.73 m2 was studied. A total of 998 patients, mean age 59.94 years were included; 112 with eGFR < 60 ml/min/1.73 m2 and 30 with eGFR < 45 ml/min/1.73 m2. Model 1 detected 816 patients: 108 with eGFR < 60 ml/min/1.73 m2 and all 30 with eGFR < 45 ml/min/1.73 m2. Model 2 detected 745 patients: 108 with eGFR < 60 ml/min/1.73 m2 and all 30 with eGFR < 45 ml/min/1.73 m2. Model 3 detected 622 patients: 100 with eGFR < 60 ml/min/1.73 m2 and all 30 with eGFR < 45 ml/min/1.73 m2. Model 4 detected 440 patients: 86 with eGFR < 60 ml/min/1.73 m2 and all 30 with eGFR < 45 ml/min/1.73 m2. Associations were significant (p < 0.001). Model 4 is most effective, resulting in the lowest proportion of superfluous eGFR measurements while detecting all patients with eGFR < 45 ml/min/1.73 m2 and most with eGFR < 60 ml/min/1.73 m2. (orig.)