Patient Safety Tip of the Week

October 25, 2011

Renewed Focus on MRI Safety



Over the last 5 years we have done a number of columns on MRI safety (see the links at the end of today’s column). But this week the FDA is conducting a workshop on MRI safety in response to ever-increasing adverse events related to MRI scanning. There has been an almost five-fold increase in MRI accidents from 2004 to 2009, from 40 reports to 193 reports in the FDA’s MAUDE database (Forrest 2011). The Forrest article on the website notes a recent review of the MAUDE incidents done by two of the most well-known researchers in the field of MRI safety, Tobias Gilk and Emanuel Kanal. They found the vast majority of incidents fell into one of three categories: (1) burns and thermal injuries (2) projectile events and (3) hearing loss, all potentially preventable events.


We’ve discussed burns from MRI in several columns (see our Patient Safety Tips of the Week for June 1, 2010 “Iatrogenic Burns” and October 5, 2010 “More Iatrogenic Burns” and our March 2009 What’s New in the Patient Safety World column “Risk of Burns during MRI Scans from Transdermal Drug Patches”).


For many years we have also considered the presence of certain devices, like cardiac pacemakers or implantable cardioverter-defibrillator devices, to be absolute contraindications to MRI scanning. Over the years there has been a focus by device manufacturers to develop devices that are “MRI-compatible”. Also, a recent study (Nazarian 2011) demonstrated a very low incidence of harm when MRI is done with appropriate precautions in patients with some select cardiac devices. We concur with the editorial accompanying that article (Reynolds 2011) that it is critical to assess the potential benefits and risks in individual patients with such devices and if the decision is to procede with MRI, all the precautions taken in the Nazarian study should be followed.


The most feared incidents in the MRI unit are the projectile events. Just published online in Patient Safety and Quality Healthcare (which, by the way, is another great free patient safety resource) is the root cause analysis of the terrible accident 10 years ago in which young Michael Colombini was fatally injured in a projectile incident in an MRI suite (Gilk 2011). The video of that root cause analysis is also available at the PROACTFor website. The video is 24 minutes long but we guarantee you it will probably be some of the best 24 minutes of your time invested. Gilk and his colleague Robert Latino have done a superb job in reconstructing all the facts in the Columbini accident from published accounts and publicly available testimony in the legal proceedings following that case.


The incident, like almost all other incidents we see with bad patient outcomes, was the result of a cascade of events, not a single error or event. And many (in fact most) of the contributory events were not temporally related to the accident but rather were related to the design, planning, and oversight of the MRI unit.


This RCA has some very valuable lessons learned. While most facilities have done a good job of inservicing their staffs about MRI dangers and safety, sometimes someone who has not been inserviced might gain access to the MRI suite. In this RCA, the two MRI technicians had gone to another room and temporarily lost auditory contact with the MRI suite, when a nurse happened to return to the MRI suite to pick something up from a prior patient encounter. That nurse heard the frantic pleas from the anesthesiologist attending the patient for oxygen. That nurse unwittingly picked up a ferromagnetic oxygen canister and handed it to the anesthesiologist. The magnet then wrested the canister from the anesthesiologist and it launched into the MRI unit, fatally injuring the patient. Apparently neither the nurse nor the anesthesiologist had ever been trained in MRI safety. The RCA goes on to discuss not only the training issues but also the fact that individuals lacking such training had relatively easy access to the MRI unit.


In our March 17, 2009 Patient Safety Tip of the Week “More on MRI Safety” we discussed many important issues related to role of the anesthesiologist in the MRI suite. We discussed the American Society of Anesthesiologists (ASA) report “Practice Advisory on Anesthetic Care for Magnetic Resonance Imaging”.  In addition to all the training related to working in the MRI suite, the anesthesiologist needs to develop a plan for implementing anesthesia care before each individual case. Such a plan should be done in collaboration with other personnel who will be involved in the care of the patient, including the MRI technician, radiologist, radiology nurse, other clinical personnel accompanying the patient and even the facility biomedical engineer. They need to find the appropriate location of moveable equipment in relation to the gauss lines within the MRI suite. The plan should include where the optimal line of sight will be for both observing the patient and any monitors. The plan needs to have a contingency for emergencies. The latter would include how to summon assistance during an emergency, where emergency medications and equipment will be located, where the patient will be evacuated to (note that rescuscitation is begun as the patient is transported to the previously designated safe area outside Zone IV), and how emergency response personnel need to be prevented from entering Zone IV. The plan also needs to address the level of sedation or anesthesia anticipated and what to do if a higher level is needed or inadvertently achieved. The plan needs to specify how the intravenous drugs, oxygen, anesthetic gases, suction and waste management will be handled. In some cases where MRI-safe or MRI-conditional equipment is not available, special conduits called “wave guides” may have to be used for some of these management activities. When deep sedation is used or when direct observation of respiration cannot be performed during moderate sedation, monitoring of exhaled carbon dioxide should be considered (simply monitoring oxygen saturation by pulse oximetry is insufficient in such circumstances). It should also be remembered that the magnetic fields may interfere with interpretation of some monitoring tools, such as EKG, so some data must be interpreted with caution.


We suggested a few additions to this otherwise excellent ASA practice advisory. First is the importance of determining up front whether the MRI is truly indicated, whether the potential benefits of performing the MRI outweigh the potential risks, and whether alternative safer imaging modalities might suffice. In many of the incidents we’ve seen occurring in ICU patients transported to the MRI suite or radiology suite, we’ve been surprised at how often the scan being done was really of marginal value.


Second is the need for a huddle/timeout before the procedure is performed. We should approach doing MRI on these critically ill patients in the same manner in which we approach patients going to the OR. A “huddle” or whatever else you’d like to call a pre-procedure briefing is very important in such cases. Not only do you need to know you have all the equipment needed, but you also need to know everyone’s role and have contingency plans for emergencies. This is where you ensure all parties know what to do if there is a fire or if there is a cardiopulmonary arrest or a “quench”. You discuss what location you will need to move the patient to in such events. You discuss the availability and location of equipment and medications you may need. You discuss the line of sight required and where the monitoring equipment will be deployed. You may need to discuss also how you will communicate (with both staff and patient) given the high noise levels associated with MRI scanning. You should probably even discuss the potential impact of the lighting levels in the various zones (and fact that you may not be able to wear your ferromagnetic glasses in Zones III and IV).


Third, really related to the above, is use of a checklist. To remember all the needs for the procedure (which vary be individual patient) and the contingencies you have to plan for is really too much to expect for any individual or group of individuals. That’s where the simple checklist comes in: it helps you to remember details you might otherwise overlook.


Fourth, you need to practice. We wonder how many MRI facilities, particularly hospital-based ones, actually simulate an emergency during MRI scanning.


There were, of course, also latent factors that contributed to the Columbini incident. One was the fact that the oxygen supply to the wall-mounted unit had no flow of oxygen at the time of the incident. Part of running an MRI or radiology unit should be regular checking of the availability of adequate oxygen supply (that should occur daily before the first patient is brought to the suite and again after every use of oxygen).


A second factor was that ferromagnetic oxygen canisters were easily accessible just outside the MRI room. That’s really an accident waiting to happen since it should be anticipated someone, in an emergent situation, might grab any available oxygen canister. In our April 8, 2008 Patient Safety Tip of the Week Print “Oxygen as a Medication” we noted a Patient Safety Advisory in 2005 “Continuity of Oxygen Therapy During Intrahospital Transport” found that as many as 55% of patients transported to radiology run out of oxygen at some time during their trip. That includes unavailability of oxygen at the destination site. If you use the “Ticket to Ride” concept for transports, ensuring adequate oxygen supply is one of the items that needs to be discussed. It is critical that if the destination is an MRI suite all staff are aware that ferromagnetic oxygen canisters cannot be taken into the MRI suite. Therefore the staff involved in the transport need to ensure that there will be adequate supply of oxygen on the receiving end (the MRI suite) either in the form of non-ferromagnetic canisters or oxygen delivered via wall outlet. Obviously in the latter the clinical staff involved in the transport should discuss with the MRI staff the status of oxygen via either of those two sources.


One of the key root causes identified was the fact that two separate entities were involved in the MRI operation and it was unclear who was providing all the safety oversight for the unit. That same issue undoubtedly still exists today where some MRI units are jointly owned by a hospital and a physician group or where a physician group owns the MRI and rents space for it from a hospital. In addition, we still see mobile MRI units. For the latter, nurses or other hospital staff typically accompany the patient to the mobile MRI unit. We wonder how many of those nurses and other staff have been inserviced on MRI safety (since the hospital itself technically does not have an MRI unit). We strongly recommend you review all aspects of safety if your staff does go on transports to such mobile MRI units.


And, of course, a major root cause was the lack of prior lessons learned. There had apparently been prior projectile events at that same facility, including one said to have been remarkably similar to the fatal accident. The hospital involved had not even been aware of those prior incidents (because technically the MRI unit was owned and operated by another entitiy). In addition, unlike incidents in a hospital-owned MRI unit, incidents from physician-owned MRI units were not reportable to the state health department. It is not known whether root cause analyses were done on those prior events, but clearly any potential lessons learned from those were not applied going forward. Similarly, all the valuable lessons in the current RCA had to wait until now to be revealed to all other sites having MRI units.


We’ve touched on only a few points brought out in this outstanding RCA. We strongly encourage you to both read the RCA done by Gilk and Latino and watch the video. This is an incredibly useful RCA full of lessons learned that you will be able to take back to your own institution.



Our prior columns on MRI safety:


February 19, 2008 Patient Safety Tip of the Week “MRI Safety

March 17, 2009 Patient Safety Tip of the Week “More on MRI Safety

September 16, 2008 Patient Safety Tip of the Week “More on Radiology as a High Risk Area

October 2008 What’s New in the Patient Safety World “Preventing Infection in MRI

March 2009 What’s New in the Patient Safety World “Risk of Burns during MRI Scans from Transdermal Drug Patches

February 1, 2011 Patient Safety Tip of the Week “MRI Safety Audit






FDA. Public Workshop – Magnetic Resonance Imaging Safety, October 25-26, 2011



Forest W. Rising number of MRI accidents prompts FDA workshop on safety. Oct. 11, 2011



Nazarian S, Hansford R, Roguin A, et al. A Prospective Evaluation of a Protocol for Magnetic Resonance Imaging of Patients With Implanted Cardiac Devices. Ann Intern Med 2011; 155: 415-424



Reynolds MR, Zimetbaum P. Magnetic Resonance Imaging and Cardiac Devices: How Safe Is Safe Enough? (Editorial). Ann Intern Med 2011; 155: 470-472



Gilk T, Latino RJ. MRI Safety 10 Years Later. What can we learn from the accident that killed Michael Colombini? Patient Safety and Quality Healthcare 2011; online first Nov-Dec 2011



PROACTFor Gilk T, Latino RJ. Movie: Colombini MRI Case: Root Cause Analysis - 10 Years Later.

PROACTFor  Michael Colombini MRI RCA - MRI Safety Week. 2011



American Society of Anesthesiologists. Practice Advisory on Anesthetic Care for Magnetic Resonance Imaging. A Report by the American Society of Anesthesiologists Task Force on Anesthetic Care for Magnetic Resonance Imaging. Anesthesiology 2009; 110(3): 459-479, March 2009



Pennsylvania Patient Safety Authority. Patient Safety Advisory.

Continuity of Oxygen Therapy During Intrahospital Transport. September 2005.















Tip of the Week Archive


What’s New in the Patient Safety World Archive