To compare ioxaglate with the traditional agent diatrizoate, 44 patients scheduled to undergo coronary angiography that included internal mammary artery (IMA) visualization were entered into a randomized double blind-parallel design protocol. Complete data were collected on 32 out of 44. Four patients were withdrawn because of angiographically normal coronary arteries; seven, because of unsuccessful IMA cannulation; and one, because of an anaphylaxis-like reaction to ioxaglate. No other serious adverse effects were seen with either agent. The major endpoints were patient and physician assessments of discomfort, rated independently on a 4-point scale. Ioxaglate caused significantly less discomfort (n = 17, median rating of "mild discomfort") than did diatrizoate (n = 15, median rating of "severe discomfort"; P less than 0.01); this effect was independent of patient sex, the number of injections, and the volume of dye injected. Radiographic quality was good with both agents. We conclude that ioxaglate is much better tolerated than diatrizoate during angiography of the IMA.