Contrast-induced nephropathy is a potentially serious untoward reaction to radiologic contrast media. The incidence of this nephropathy and the predisposing conditions are not well established, possibly because of methodologic differences between studies. We evaluated the incidence of contrast-induced nephropathy after femoral arteriography in 394 patients by using multiple definitions (different increases in serum creatinine or blood urea nitrogen levels at various times). When an increase in the level of serum creatinine of greater than 0.3 mg/dl and greater than 20% on day 1, 2, or 3 and on day 5, 6, or 7 was used to define the disorder, the incidence in our group of patients was 10% for nonazotemic patients vs 30% for azotemic patients (p less than .001); 2% for nondiabetic, nonazotemic patients vs 16% for diabetic, nonazotemic patients (p = .003); and 38% for patients who were both diabetic and azotemic vs 16% for diabetic, nonazotemic patients (p = .022). Baseline renal insufficiency and diabetes mellitus (especially when insulin dependent) were significant predisposing factors. The effects of dehydration and increased volume of contrast medium on the incidence of contrast-induced nephropathy were not clear; the age and sex of the patient were not important risk factors. The incidence of contrast-induced nephropathy depends on the definition used. Although contrast-induced nephropathy may develop in any patient, diabetes, renal insufficiency, and, possibly, dehydration and dose of contrast medium are risk factors.