Three hundred seven high risk patients with renal impairment (serum creatinine greater than or equal to 1.5 mg/dl) were randomized in a double-blind manner to either iopamidol (a nonionic, low osmolar radiocontrast agent) or diatrizoate (a conventional radiocontrast agent) at cardiac angiography with subsequent follow-up study of renal function. Baseline clinical and angiographic variables were similar in the iopamidol (n = 155) and diatrizoate (n = 152) groups. Change in renal function after angiography was less pronounced with iopamidol compared with diatrizoate as measured by mean ( +/- SD) increase in 24 h serum creatinine (0.11 +/- 0.2 versus 0.22 +/- 0.26 mg/dl, p less than 0.001), mean maximal increase in serum creatinine (0.2 +/- 0.44 versus 0.38 +/- 0.73 mg/dl, p less than 0.0001) and percent of patients with a maximal increase in serum creatinine greater than 0.5 mg/dl (8% versus 19%, p less than 0.01). Such differences could not be documented in diabetic patients using insulin. There was no significant difference between agents in the number of patients developing clinically severe acute renal dysfunction. It is concluded that iopamidol is less nephrotoxic than diatrizoate in high risk patients at cardiac angiography. However, the difference in nephrotoxicity is small, of no major clinical significance in the majority of high risk patients and could not be documented in insulin-using diabetic patients. Iopamidol may be the preferred agent in certain patients with advanced renal impairment, but further study is warranted.