The aim of this study was to describe our experience and institutional savings with a selective use of low-osmolality contrast media (LOCM) in CT. From 1995 to 1998, a total of 19,834 contrast-enhanced CT examinations were performed at our institution. Contrast was injected with a power injector and large venous catheter, 20-G for 2- to 3-ml rate and 18-G for 3- to 5-ml rate. High-osmolality contrast media was used in 13,670 patients (71%). The LOCM was used in 5884 (29%) patients. Our guidelines for the use of LOCM included cardiac dysfunction, severe pulmonary impairment, history of allergy or prior moderate reaction to HOCM and severe debilitation. Prior to the injection of HOCM, 10 mg of metoclopramide (Primperan, Delagrange Quétigny, France) were administered to reduce nausea and vomiting. In the HOCM group there were 304 minor or mild adverse reactions (2.2%), and 10 severe adverse reactions (0.08%). In the LOCM there were 34 mild or moderate adverse reactions (0.59%) and 3 severe adverse reactions (0.05%). Significant differences in terms of mild adverse reactions were found between HOCM and LOCM (Fischer's test, p<0.001). No significant differences were found in terms of severe adverse reactions ( p=0.27). After subtracting the cost of treating additional adverse reactions, the net differential cost between universal and selective use of LOCM was 565,285 Euro (601,067 US dollars). This means a net increase of 41.4 Euro per patient or 414,000 Euro per 10,000 patients (438,840 US dollars). Selective use of LOCM in CT is safe and effective and results in a substantial reduction in costs.