RATIONALE AND OBJECTIVES. The authors evaluated the image quality on delayed CT arthrography images with the use of the nonionic dimeric contrast agent, iodixanol 270 mg I/mL, compared with iohexol 300 mg I/mL, METHODS. A total of 132 patients with shoulder pain were included in a randomized, parallel, double-blind study. Sixty seven patients received iodixanol and 65 patients received iohexol, Patients underwent two CT-arthrography examinations: the first was performed 20 to 30 minutes after contrast injection and the second, 50 to 70 minutes after contrast injection. Data from 31 patients were excluded from the efficacy analysis. The overall quality of CT arthrography images was graded into four categories: excellent, good, moderate, and bad. RESULTS. The overall quality of delayed CT arthrography images was significantly better in the iodixanol group (P = 0.001, alpha = 0.05), On the first CT examination, image quality was good or excellent in 88% of the cases in the iodixanol group and in 96.1% in the iohexol group. The results on the delayed CT arthrography examination indicated that image quality was good or excellent in 88% of the cases when iodixanol was used and in 63.5% when iohexol was used. CONCLUSIONS. The overall quality of delayed images nas significantly better with iodixanol than with iohexol.