Objective: To compare carbon dioxide-digital subtraction arteriographic (CO2-DSA) images of renal artery anatomy with standard iodinated contrast arteriographic (ICA) images.
Design: One hundred patients with vascular disease who required abdominal aortography were evaluated by both CO2-DSA and ICA modalities. Two blinded readers interpreted arteriograms for the degree of renal artery stenosis, and a third reader was employed to resolve differences in reader interpretations.
Setting: University medical center.
Main outcome measures: The sensitivity, specificity, negative predictive value, positive predictive value, and accuracy were calculated for the ability of CO2-DSA to demonstrate a 60% or greater stenosis of the main renal artery; kappa values for CO2-DSA and ICA were calculated to assess intraobserver variability.
Results: Of the 200 main renal arteries imaged, 17 (9 by means of CO2-DSA), 8 means of ICA) were eliminated because of inadequate visualization of the renal artery. In identifying a renal artery stenosis of 60% or greater, CO2-DSA had a sensitivity of 0.83, specificity of 0.99, positive predictive value of 0.94, and negative predictive value of 0.98. The overall accuracy was 0.97. The kappa was 0.75 for CO2-DSA and 0.70 for ICA, hence, the variation in the interpretations of CO2-DSA and ICA were comparable.
Conclusion: Images by means of CO2-DSA accurately reflect pathologic changes in renal arteries and are thus useful in the diagnosis of clinically occult occlusive renal artery disease in patients at risk of contrast medium-related nephrotoxicity.