By Joshua D. Moss, MD

Associate Professor of Clinical Medicine, Cardiac Electrophysiology, Division of Cardiology, University of California, San Francisco

Dr. Moss reports he is a consultant for Biosense Webster and Abbott.

SYNOPSIS: Performing an MRI on patients with pacemakers and defibrillators that were not MRI-conditional was not associated with any clinically significant changes in device parameters.

SOURCE: Shah AD, Morris MA, Hirsh DS, et al. Magnetic resonance imaging safety in nonconditional pacemaker and defibrillator recipients: A meta-analysis and systematic review. Heart Rhythm 2018;15:1001-1008.

The major cardiac implantable electronic device (CIED) companies have engineered and tested pacemaker and implantable cardioverter-defibrillator (ICD) pulse generator and lead systems with demonstrated safety in MRI scanners, classified as MRI-conditional devices. However, performing an MRI on a patient with a nonconditional device remains relatively contraindicated over concerns regarding lead malfunction, battery depletion, device damage, and inappropriate shock. Scanning of patients with nonconditional pacemakers and defibrillators has been limited to tertiary academic centers that follow specific MRI protocols. Although studies have shown these protocols to be safe, without significant adverse events, performing an MRI with a nonconditional CIED is not covered by the Centers for Medicare & Medicaid Services (CMS) due to safety concerns.

Shah et al sought to further characterize the clinical risks associated with performing an MRI on nonconditional CIED recipients by conducting a meta-analysis of 70 observational studies. More than 5,000 patients with nonconditional devices who underwent MRI were included (3,692 patients with a pacemaker, 1,440 patients with an ICD). Roughly 10% of all patients were reported to be pacemaker-dependent. The brain and cervical spine were imaged most frequently. Most MRI scanners used a ≤ 1.5 Tesla magnet. In all but six observational studies, one of several periprocedural MRI protocols was followed, including the American Heart Association 2007 protocol and the John Hopkins MRI protocol. In older studies, various therapies simply were disabled. Notably, there were several scenarios included that often are avoided, even with MRI-conditional devices. These included scans of 100 abandoned leads, a small number of externalized leads connected to pulse generators taped to the chest wall, very recent implants, and thoracic isocenter imaging.

Statistically significant changes in lead parameters, such as sensing and pacing impedance, were observed, but none were clinically significant. There also was a statistically significant but clinically insignificant acute decrease in battery voltage of 0.002 ± 0.001 V. There were three lead failures, none of which were clearly directly attributable to the MRI study. There was one inappropriate ICD shock, which occurred after an MRI was performed without knowledge that the patient had an implantable defibrillator. Electrical reset to factory settings was observed in 1.6% of scans, but only in devices that were implanted before 2006. Symptoms possibly attributable to MRI-induced torque, current induction, or lead heating were reported in 0.3% of scans. There was an increased risk of a safety event when a > 1.5 Tesla MRI magnet was used, driven by power or device resets. The authors concluded that clinical events after MRI of nonconditional CIEDs are rare, and that such scans can be performed safely in appropriately screened patients.


Although MRI-conditional CIEDs are increasingly common in clinical practice, nonconditional devices still are encountered frequently in both inpatient and outpatient settings. It is estimated that a large number of these patients will warrant an MRI in their lifetime. Concerns about performing an MRI on patients with nonconditional devices arise from both observed and theoretical interactions between the device and the surrounding magnetic field emitted by the MRI. The dynamic nature of MRI-induced magnetic fields can result in heating of the myocardium via the implanted lead, inappropriate sensing of ambient signals leading to inhibition of pacing or inappropriate defibrillation, and hardware resets or damage within the pulse generator. As a result, guideline and regulatory statements have recommended against MRI scans in those with a nonconditional CIED, even when an MRI may be the imaging modality of choice. Several protocols have been developed to optimize safety, including the American Heart Association 2007 protocol and the Johns Hopkins MRI protocol. The authors of large studies have tested these protocols and found a low rate of clinical events during and after MRI. Still, scanning nonconditional devices has not been covered by CMS or adopted widely.

This well-conducted meta-analysis adds to the growing body of evidence that performing an MRI on patients with nonconditional CIEDs is safe and only rarely associated with adverse clinical events. It is important to note that most patients who underwent MRI were managed with a specific MRI protocol, including periprocedural device interrogation and reprogramming, cardiac telemetry or plethysmography monitoring during the scan, and verbal contact during the entirety of the scan. An electrophysiologist or CIED-trained nurse or technician also was available immediately. Patients with unique situations, including abandoned leads, epicardial leads, and very recently implanted devices (< 24 hours), also were included without any significant events. However, they were not well represented, and safety results should not necessarily be extrapolated to these unique scenarios.

In our own practice, patients with nonconditional devices are scanned according to an evidence-based protocol that includes pre- and post-MRI device interrogation with reprogramming (Learn more at: Careful screening of these patients remains critical, as data on the overall safety of scanning patients with abandoned leads, externalized leads, and very recently implanted devices remain limited.