Experimental studies have suggested that nonionic contrast agents are less nephrotoxic than ionic contrast agents. To examine the relative nephrotoxicity of the two types of agents, we randomly assigned 443 patients to receive either iopamidol (nonionic) or diatrizoate (ionic) for cardiac catheterization. The patients were stratified into low-risk (n = 283) or high-risk (n = 160) groups, on the basis of the presence of diabetes mellitus, heart failure, or preexisting renal insufficiency (base-line serum creatinine level, greater than 133 mumol per liter). Serum and urine analyses were performed at base line and 24 and 48 hours after the infusion of contrast material. Nephrotoxicity was defined as an increase in the serum creatinine level within 48 hours of at least 44 mumol per liter. The median maximal rise in the serum creatinine level was 18 mumol per liter in both the diatrizoate group (n = 235) and the iopamidol group (n = 208) (P not significant; power to detect a difference greater than 9 mumol per liter, greater than 90 percent). Creatinine levels increased by at least 44 mumol per liter (0.5 mg per deciliter) in 10.2 percent of the patients receiving diatrizoate and 8.2 percent of the patients receiving iopamidol (P not significant). Among the high-risk patients, creatinine levels increased by at least 44 mumol per liter in 17 percent of the patients in the diatrizoate group, as compared with 15 percent of the patients in the iopamidol group (P not significant). We were unable to demonstrate a difference in the incidence of nephrotoxicity between patients receiving a non-ionic contrast agent and those receiving an ionic contrast agent.