Safety of MRI in Patients With Cardiac Devices

Study Questions:

Can magnetic resonance imaging (MRI) be safely performed in patients with legacy pacemakers or implantable cardioverter-defibrillators (ICDs)?


The authors conducted a prospective nonrandomized study evaluating the accuracy and safety of 1.5 Tesla MRI in patients with pacemakers and ICDs, which are not MRI-conditional, at a single tertiary center. Patients with recently implanted leads (<4 weeks), epicardial leads, nonfunctional leads, subcutaneous ICD systems, or who were pacemaker dependent or had ICDs without an asynchronous pacing mode were excluded. Device parameters were measured prior to and minutes after the MRI. Programming was changed to asynchronous pacing for those with intrinsic heart rates <40 bpm and an inhibited mode for all others. Follow-up device interrogation was recommended at 6 months. MRI sequences were modified to limit specific absorption rates to <2.0 W/kg in the first 55 patients, but were not restricted in subsequent patients. Prespecified outcomes included generator failure, power-on reset, changes in pacing threshold or sensing requiring system revision or reprogramming, battery depletion, cardiac arrhythmia, inhibition of pacing, and inappropriate delivery of therapies such as tachycardia pacing or shock. Patient reported symptoms such as warmth at the device, and palpitations were also recorded. In patients who returned, lead and generator parameters were again recorded as no change (≤20%), expected change (20-50%), or notable change (>50%).


A total of 2,130 MRI scans were performed in 1,509 patients, of whom 880 (58%) had pacemakers and the remainder had ICDs. Adverse events were rare. Power-on reset occurred in 9 MRI scans (0.4%), but in most cases (3 out of 8 patients), the examination could still be completed. Five other MRI scans were terminated for reasons including functional inhibition of pacing, frequent nonsustained ventricular tachycardia, and extensive image artifact. No changes in device parameters large enough to result in lead or system revision or reprogramming occurred. However, 1% of patients had >50% reduction in P-wave amplitude acutely and 4% at long-term follow-up. Similar percentages had changes in atrial and ventricular capture thresholds. ICDs had greater changes in sensing and battery voltage than pacemakers. Changes in sensing and capture thresholds were more common with longer leads and after multiple MRIs. No association was noted between thoracic and nonischemic imaging and changes in device parameters.


Patients with legacy ICDs and pacemakers may safely undergo MRI examinations.


The Johns Hopkins group has pioneered MRI in patients with ICDs and pacemakers and now presents a very large registry showing the safety of their protocol. They have previously published similar results on the initial 68 and 522 MRI scans. The recently published multicenter MagnaSafe registry data (n = 1,000) also support this, as well as smaller single-center studies from other sites. It is nonetheless important to recognize that these scans require input from electrophysiology specialists and careful consideration of risks and benefits. Repeated scanning may have greater consequences than single imaging sessions.

Clinical Topics: Arrhythmias and Clinical EP, Noninvasive Imaging, Implantable Devices, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias, Magnetic Resonance Imaging

Keywords: Arrhythmias, Cardiac, Artifacts, Defibrillators, Implantable, Diagnostic Imaging, Electrophysiology, Magnetic Resonance Imaging, Pacemaker, Artificial, Tachycardia, Ventricular

< Back to Listings