Risk of Nephropathy Is Not Increased by the Administration of Larger Volume of Contrast During Coronary Angiography

Abstract

It is proposed that contrast-induced nephropathy (CN) correlates with the use of increasing contrast volumes during coronary angiography. This supposition has led to the current recommendation to limit the dose of contrast in patients at high risk for renal dysfunction. Limits in contrast dosing may negatively impact the evaluation of patients undergoing cardiac catheterization for myocardial infarction and acute coronary syndrome. The objective of this study was to empirically assess, in a large population, the presence and strength of this correlation. Baseline blood samples and clinical information were obtained from 5256 consenting patients hospitalized for coronary angiography. Levels of serum creatinine were measured pre- and postcatheterization, and the total change in serum creatinine was calculated. Nephropathy was defined as a change of ≥0.5 mg/dL. The total volume of contrast dye (iopamidol, nonionic) used during the angiography procedure was recorded. Logistic regression was used for the primary analysis. The average age was 64 ± 14 years, and 67% of patients were male. Paradoxically, the incidence of CN was inversely related to the volume in the overall population: 16%, 14%, 8%, and 7% for quartile (Q) 1 (225 mL) of contrast, respectively (P-trend P < 0.001). Other predictors included age, left ventricular ejection fraction, diabetes, and baseline creatinine level (all P < 0.001). For patients at high risk, with a baseline creatinine >2.0 mg/dL (n = 415), contrast volume (Q1: 170 mL) did not predict either increased or decreased risk of CN (48%, 42%, 49%, 43%, respectively, P-trend = 0.76). This lack of predictive value remained after multivariable adjustment. In this large population, no association was found between the amount of contrast used during angiography and the incidence of CN in patients at initial high risk. The apparent inverse relation of risk with volume in the overall population is likely explained by clinical practice bias. If confirmed, these results may have important clinical implications.