Members of the Society of Cardiovascular and Interventional Radiology and the Society of Uroradiology were surveyed to assess both current practice and determinants of use of low-osmolality contrast agents (LOCA). Of 715 questionnaires mailed, 229 were returned (32%). Many indicated that the responses referred to the practice of a group, rather than to an individual. Two thirds of the responses indicated that the choice of agent was made by the radiologist alone; one third shared the decision with administrators. Approximately 40% of respondents use written guidelines, but this percentage was higher when administrators were involved in the decision. Almost all (99%) obtained written informed consent--one third for all procedures that use contrast material and two thirds for angiographic studies only. Four percent noted the availability of LOCA in obtaining consent. Overall, LOCA were used for most angiographic procedures by 38% and for most IV studies by only 9% of respondents. All used them for at least some angiographic studies, and all but 2% used them for some IV studies. The major indications for use in both angiographic and IV studies were a prior major (but not minor) reaction and renal dysfunction. Anticipated patient discomfort also was important for intraarterial use. LOCA were used most often for peripheral and pulmonary arteriography. When LOCA were used, nonionic agents were preferred to low-osmolality ionic dimeric contrast agents for coronary angiography and venography, but this preference was essentially reversed for peripheral arteriography. This survey suggests that IV use of LOCA is much less frequent than intraarterial use and that use is individualized to certain risk factors and certain examinations. Guidelines for use are often in place, but this is clearly not universal. The importance of informed consent for use of contrast agents and of the appropriate role for LOCA remains unsettled.