Table 1.

Studies reporting MR safety from scanning patients with a CIED.

First Author, Publication Year, Country	Population Size City, Country	Study Design Institution (s)	Cardiac Implantable Electronic Devices	Field Strength Sequences	MRI Scans Anatomic Regions	Outcomes	Findings
Nazarian, et al.[9] 2011 United States and Israel	438 patients 555 MRIs Baltimore, Maryland, USA and Haifa, Israel	2 center prospective non-randomized trial Johns Hopkins University, USA and Rambam Medical Center, Israel	Non-MR-conditional PPM (n = 237 patients) or ICD (n = 201 patients) Excluded abandoned or epicardial leads	1.5T Standard sequences	•89 cardiac •344 brain/spine •79 abdomen/pelvis •50 extremity	•Activation or inhibition of pacing •Symptoms •Device variables	•3 patients (0.7%) had a power-on reset, although without device dysfunction during long-term follow-up •RV sensing and atrial and right and left ventricular lead impedances were reduced immediately post-MRI •At long-term follow-up, there was decreased RV sensing, decreased RV lead impedance, increased RV capture threshold and decreased battery voltage •None required device revision /reprogramming
Camacho, et al.[33] 2016 United States	104 patients 113 MRIs Atlanta, Georgia, USA	Single center retrospective cohort study Emory University, USA Dates of the scans were not provided	Non-MR-conditional PPM (n = 74 scans) or ICD (n = 39 scans) Abandoned or capped leads were excluded 5 patients were pacer dependent	1.5T Standard sequences	•3 cardiac •5 chest •47 abdomen/pelvis •81 brain, C/T/L spine •3 neck •2 extremity	•Changes in lead impedance, sensing, or thresholds •Episodes of electromagnetic interference or noise •Programming changes before or after the MRI •Patient symptoms •Abnormal device activity	•No significant changes in lead parameters •Electromagnetic noise was detected on at least 1 lead in 7.1% of studies •Patients reported transient symptoms during 3 examinations (heating at the pocket site, tingling at the pocket site, and palpitations) without complications •No abnormal device activity •No emergency termination of the MRI •All studies were diagnostic
Nazarian, et al.[13] 2017 United States	1509 patients 2103 MRIs Baltimore, Maryland, USA	Single center prospective observational cohort study Johns Hopkins University, USA February 2003Through January 2015	Non-MR-conditional PPM (n = 880 patients) or ICD (n = 629 patients) 137 patients with device dependence. Pacers year 1996 or later and ICDs year 2000 or later were included	1.5T Standard sequences	No cardiac scans •257 thoracic	•Generator failure •Power-on reset •Change in pacing or sensing thresholds requiring programming changes •Battery depletion •Cardiac arrhythmia •Inhibition of pacing •Inappropriate ATP or shock •Patient symptoms	•9 MRI scans (0.4%) in 8 patients (0.5%) had power-on reset; this was transient in all but 1 scan •Devices were manufactured between 1997-2009 -Device program failure in 1 device (less than 1 month of battery remaining) •-No long-term clinically significant events
Russo, et al.[12] 2017 United States	1246 patients 1500 MRIs Multiple locations in the USA	Multi-center prospective observational cohort study April 2009 through April 2014 Scripps Research Institute, USA and 19 centers in the USA	Non-MR-conditional PPM (n = 818 patients) or ICD (n = 428 patients) 1000 pacer scans 500 ICD scans	1.5T Standard sequences	No cardiac scans •591 brain •249 C spine •448 L spine •168 extremity/joint •81 abdomen/pelvis •172 other	•Death •Generator or lead failure requiring immediate replacement •Loss of capture (for pacer dependent patients) •New arrhythmia •Partial or full generator electrical reset	•Nodeaths •No ventricular arrhythmias •No lead failure •6 cases of self-terminating atrial fibrillation/flutter •6 cases of partial electrical reset •1 ICD device programming failure due to protocol violation
Okamura, at al.[29] 2017 United States	9 patients with PPM and ICD with a nearly depleted battery 13 MRIs Rochester, Minnesota, USA	Single center retrospective observational cohort study Mayo Clinic, USA January 2008 to May 2015	8 scans with devices at ERI Non-MR-conditional PPM (n = 4 scans with a device with a nearly depleted battery)) or ICD (n = 9 scans with a device with a nearly depleted battery) Pacer dependent patients were excluded	1.5T Standard sequences	No cardiac scans •11 head •3 chest Some patients had both scans at the same time	•Power on reset •Elective replacement indicator (ERI) turned on •Unable to reprogram the device •All events occurred in pacemakers implanted before 2005	•2 scans with pacers close to ERI resulted in a power on reset •1 scan with a pacer close to ERI resulted in a power on reset during MRI and automatically changed to VVI mode •1 scan with a pacer at ERI did not allow reprogramming
Do, et al.[32] 2018 United States	111 patients 111 MRIs Los Angeles, California, USA	Single center retrospective observational cohort study UCLA, USA April 2013To October 2016	Non-MR-conditional PPM (n = 12 patients), ICD (n = 73 patients), and CRT-D (n = 29 patients) 3 patients were device dependent (1 with pacer, 1 with ICD, and 1 with CRT-D) Out of 114 consecutive studies, 3 scans were stopped prematurely and excluded due to anginal chest pain, anxiety, and frequent non-sustaiend VT prior to the scan	1.5T Wideband sequences for late gadolinium enhancement (LGE)	Cardiac scans •111 cardiac	•Clinical deterioration nor death during the scan •Generator failure requiring replacement •Lead failure requiring replacement •New-onset arrhythmia •Loss of capture in pacemaker-dependent patients Power-on reset	•No adverse CIED complications or clinical outcomes •87% had no artifact limiting interpretation
Shah, et al.[135] 2018 United States	5099 patients 5908 MRIs Multiple locations	Systematic review & meta-analysis 70 studies were included in the systematic review 31 studies were included in the meta-analysis cohort	Non-MR-conditional devices (3147 RA leads, 4023 RV leads, 268 LV leads); 1440 defibrillator leads; 100 abandoned leads, 25 epicardial leads, 4 subcutaneous ICD, small number of temporary pacemakers 3692 pacer patients 1440 ICD patients, 268 LV pacing leads 551 pacer dependent patients 39 patients with AICD and device dependent	0.2T 0.5% 1.5T 2T 3T Standard sequences	No cardiac scans •773Thoracic •3105 head/neck •1153 abdomen/pelvis/L spine •402 extremity	•Deaths •Lead survival •Lead performance •Electrical reset •Inappropriate ICD shock and therapy •High voltage impedance •Patient symptoms •Battery voltage change	•No deaths1 ICD shock (inadvertently scanned at 0.2 T) •3 Lead failures (none directly attributable to MRI) •94 electrical resets (all devices older than 2006) •No changes in lead parameters, battery or generator performance
Lupo, et al.[27] 2018 Italy	120 patients 142 MRIs Milan, Italy	Single center prospective cohort study Humanitas University, Italy December 2006 to November 2014	Non-MR-conditional PPM (n = 71 scans) or ICD (n = 71 scans) Pacer dependent patients were excluded No abandoned or epicardial leads	1.5T Standard sequences	Cardiac scans •55 cardiac •60 brain, C/T/L Spine •7 thoracic •3 vascular	Primary: Frequency of adverse events within 3 h after the MRI scan •Requiring life-support procedures •Not requiring life-support procedures •Device modification •Any other adverse event related or unrelated to MRI •Secondary: increase in blood markers and rate of adverse events at follow-up (myoglobin, myocardial band isozyme, troponin)	•No adverse events •No device malfunctions •No significant changes in markers of myocardial necrosis
Padmanabhan, at al.[30] 2018 United States	80 patients with abandoned leads 97 scans Rochester, Minnesota, USA	Single center retrospective observational cohort study Mayo Clinic, USA January 2008 to March 2017	Abandoned leads ONLY Non-MR-conditional PPM (n = 31 patients) or ICD (n = 19 patients) or CRT-D (n = 13 patients) or CRT-P (n = 2 patients) or no device (n = 15 patients)) 10 patients with epicardial leads 4 patients with fragmented leads	1.5T Standard sequences	No cardiac scans •38 head •22 chest •29 lumbar •8 extremity	Primary endpoint: •Difference in adverse event rate and post-MRI serum cTnT value bewteen the study cohort and control group •Death •Generator failure •Lead failure •Loss of capture •Observed atrial arrhythmia •Ventricular arrhythmia •Electrical reset Secondary endpoints Adverse events in the performance of MRI Significant change in device parameters post-MRI cTnT values pre- and post-MRI •Contribution of body part scanned, number of ICD coils, and multiple MRI scans on cTnT values	•No adverse events •No evidence of myocardial injury (cTnT)
Nyotowidjojo, et al.[31] 2018 United States	238 patients 339 MRIs Tucson, Arizona, USA	Single center retrospective observational cohort study University of Arizona, USA December 2013To July 2016	Non-MR-conditional PPM (n = 111 patients) or ICD (n = 89 patients) or CRT-P (n = 2 patients) or CRT-D (n = 36 patients) Abandoned leads (n = 6 patients) Epicardial leads (n = 7 patients)	1.5T Standard sequences	Cardiac scans •73 cardiac •8 chest •240 non-thoracic	•Adverse clinical outcomes •Arrhythmias Patient reported symptoms	•1 full power on reset (patient with a CRT-D device which was reprogrammed successfully) •No adverse CIED complications or clinical outcomes •No significant differences between thoracic and non-thoracic scans
Vuorinen, et al.[28] 2019 Finland	793 patients 1000 MRIs Helsinki, Finland	Single center retrospective cohort study University of Helsinki, Finland November 2011 to April 2017	Non-MR-conditional PPM (n = 739 scans); ICD (n = 45 scans); CRT-D (n = 31 scans)’ CRT-P (n = 0 scans) All devices except one were implanted in 2003 or later 22 scans in 17 patients with abandoned leads, including 1 patient with an abandoned epicardial pacing lead	1.5T Standard sequences	Cardiac scans •144 cardiac •555 head/spine •15 thoracic •200 abdomen/pelvis •131 extremity/joint •12 other	•Generator failure •Power-on reset •Clinically relevant changes in pacing threshold or sensing requiring system revision or programming changes •Unexpected battery depletion •Inhibition of pacing •Patient reported events	•1 pacer dependent patient fell into elective replacement indicator (ERI) mode due to a temporarily programmed high output voltage. •1 non-pacer dependent patient had the device fall into full electrical reset mode due to electromagnetic interference (later reprogrammed without issues). •1 patient had a noise reversion notification which was later reprogrammed without issues.
Gupta, et al.[26] 2020 United States	532 patients 608 MRIs Falls Church, Virginia, USA	Single center prospective observational cohort study INOVA Heart and Vascular and Virginia Heart, USA September 2015 to June 2019	Non-MR-conditional devices (279 pacemakers; 184 ICDs; 26 CRT-P; 105 CRT-D; 2 subcutaneous ICD; 1 hemodynamic monitor; 25 scans with abandoned leads) •121 pacemaker dependent patients •43 ICD and device dependent patients •14 CRT-D and device dependent patients	1.5T Standard sequences	Cardiac scans •69 cardiac •174 head •221 C/T/L spine •22 hip/pelvis sacrum •21 shoulder •30 knee •49 abdomen •22 other	•Lead impedance change •Lead sensing change •Lead threshold change •Battery voltage change •Rhythm changes •Oxygen desaturation •Heart rate changes •Blood pressure changes •Patient symptoms •Syncope •Cardiac arrest •Death	•1 Patient with transient change in lead impedance with return to baseline •No patient events
Schaller, et al.[41] 2021 United States	139 patients 200 MRI scans with at least one abandoned leads Philadelphia, Pennsylvania, USA	Single center retrospective observational cohort study University of Pennsylvania, USA January 2013To June 2020	Active devices with abandoned leads: •51 single and dual chamber pacers •81 single and dual chamber ICDs •61 biventricular pacers/ICDs •4 subcutaneous ICDs •3 other devices Abandoned leads •55 right atrial •172 right ventricular •6 coronary sinus •4 left ventricular •5 lead fragments •1 subcutaneous array 64 patients were pacer dependent	1.5T Standard sequences	Cardiac scans •50 cardiac •1 chest •140 face/orbit/ head/brain/neck and C/T/L spine •15 abdomen/pelvis and rectum •4 prostate •9 shoulder, knee, hip, foot, ankle	•Variation in pre- and post-MRI capture threshold of 50% or more, sensing 40% or more, and lead impedance of 30% or more •Burning or pulling sensations in the chest or device pocket •Sustained tachyarrhythmias during MRI •Changes in vital signs attributable to MRI-related programming changes •Power-on resets •Change in pacing rate	•No abnormal vital signs or sustained tachyarrhythmias •No changes in battery voltage, power-on reset events, or changes of pacing rate •Transient decrease in right atrial sensing in 4 patients •Transient decrease in left ventricular R wave amplitude in 1 patient •Sternal heating resolved with premature cessation in 1 patient with an abandoned subcutaneous array
Bhuva, et al.[35] 2022 United Kingdom and United States	970 patients 1148 MRIs 615 scans with non-MR conditional systems 111 MRI scans with mismatched CIED-lead vendors; 105 MRI scans with abandoned, epicardial, or very old leads (pre 2001), or scanned < 6 weeks post implant 533 scans with MR conditional systems London, UK and Philadelphia, Pennsylvania, USA	Multi-center prospective (Barts Heart Center, UK and University of Pennsylvania, USA) and retrospective (Royal Brompton Hospital, UK) cohort study 2014 and 2019	Non-MR-conditional PPM (n = 330 scans), ICD (n = 168 scans), CRT-P (n = 26 scans), and CRT-D (n = 91 scans) MR-conditional PPM (n = 332 scans), ICD (n = 149 scans), CRT-P (n = 15 scans), and CRT-D (n = 37 scans) Abandoned leads, permanent epicardial lead,devices manufactured prior to 2001, were included as non-MR conditional scans	1.5T Standard sequences	Cardiac scans Non-MR-conditional devices: •185 cardiac •158 spine •202 head •91 abdomen/pelvis •26 extremity/joint •3 other MR conditional devices •321 cardiac •122 spine •93 head •46 abdomen/pelvis •9 extremity/joint •6 other	•Death, lead failure, sustained symptomatic or life-threatening arrhythmia, complete or partial electricalreset, generator malfunction, inappropriate inhibition of pacing, or inappropriate anti-tachycardia therapies.	•2 safety events with non-MR conditional devices •1 scan with an inaccurate battery status fault code requiring generator change (generator was already under a manufacturer advisory) •1 scan where MRI not performed due to tachycardia on scan initiation •No deaths or lead failure •No complete or partial electrical resets •No inappropriate inhibition of pacing •No inappropriate anti-tachycardia therapies during or immediately after MRI
Fluschnik, et al.[34] 2022 Germany	97 patients 132 MRI scans Hamburg, Germany	Single center retrospective cohort study April 2020 to May 2022	Non-conditional devices (n = 35 scans, including 11 scans with pacer dependent patients) Conditional devices (n = 97 scans, including 15 scans with pacer dependent patients)	3T Standard sequences	Cardiac scans Non-MR conditional devices: •2 cardiac •2 thoracic •23 head •4 abdomen/pelvis •2 whole spine/aorta •4 cervical/ lumbar spine MR Conditional devices: •2 cardiac •54 head •16 abdomen/pelvis •11 whole spine/aorta •10 cervical/lumbar spine •2 extremity/joint	•All-cause death •Arrhythmias •Loss of capture •Inappropriate anti-tachycardia therapies •Electrical reset Lead or generator failure during or shortly after MRI	•No adverse events occurred during or shortly after MRI