Purpose: To determine the success rate, safety and complications using a standard protocol and trained nurses to provide sedation for MRI under the supervision of a radiologist. Materials and Methods: Nurses were trained to provide sedation via a standard protocol for pediatric patients undergoing diagnostic MRI. Oral chloral hydrate (80-100 mg·kg-1) was used for children less than 18 mo of age. Older children received either 1-6 mg·kg-1 pentobarbital iv, with or without 1-2μg·kg·hr-1 fentanyl, or 25 mg·kg-1 thiopental pr. Sedation was defined as successful if it allowed completion of the MRI without image distorting patient movement. The records of 572 MRIs performed on 488 pediatric patients (mean age 5 ± 4 yr; age 2 mo- 14 yr) from 1991 to July 1995 were reviewed to determine the success rate and complications using the sedation program. Results: Most, 91.8% (525/572), of the MRIs were successfully completed in 445 patients. The reasons for failure were inadequate sedation (45, 95.7%) and coughing (2, 4.2%). The failure rate was much higher before 1994 (38/272, 14%) than after (9/300, 3%; P < 0.0001). Failure was more common if rectal thiopental was used (23/172, 14%) than intravenous pentobarbital (19/256, 7.4%; P < 0.05). The failure rate was also high in patients with a history of a behavioural disorder (10/59, 17%). There were no deaths or unexpected admissions as a result of the sedation program. Conclusion: A high success rate can be achieved as experience is gained using a standard protocol and trained nurses to sedate children for MRI.