Author + information
- Edward T Martin, MS, MD, FACC*,†,* (, )
- James A Coman, MD, FACC*,†,
- Frank G Shellock, PhD‡,
- Christopher C Pulling, MS§,
- Robert Fair, ARRT(R) (MR)* and
- Kim Jenkins, ARRT(R) (MR)*
- ↵*Reprint requests and correspondence:
Dr. Edward T. Martin, Oklahoma Heart Institute, 9228 South Mingo Road, Tulsa, Oklahoma 74133, USA.
Objectives The study was done to determine whether patients with pacemakers could safely undergo magnetic resonance imaging (MRI) at 1.5-Tesla (T).
Background Because of theoretical risks, it is an absolute contraindication for a patient with a pacemaker to undergo MRI. However, there are times when an MRI is needed to provide valuable clinical information.
Methods Fifty-four patients underwent a total of 62 MRI examinations at 1.5-T. The type of MRI examination was not limited and included cardiac, vascular, and general MRI studies using various whole-body averaged specific absorption rate (SAR) of radiofrequency power. Restrictions were not placed on the type of pacemaker present in the patient. All pacemakers were interrogated immediately before and after MRI scanning, and patients were continuously monitored. Before and after MRI, interrogation was done, and pacing and sensing thresholds, as well as lead impedances, were all measured.
Results A total of 107 leads and 61 pulse generators were evaluated. No adverse events occurred. Forty (37%) of the leads underwent changes, whereas 10 (9.4%) leads underwent a significant change. Only 2 of the 107 (1.9%) leads required a change in programmed output. Threshold changes were unrelated to cardiac chamber, anatomical location, peak SAR, and time from lead implant to the MRI examination. Electrocardiographic changes and patient symptoms were minor and did not require cessation of MRI.
Conclusions Safety was demonstrated in this series of patients with pacemakers at 1.5-T.
Permanent cardiac pacemakers have represented a contraindication to magnetic resonance imaging (MRI) (1,2). Strong static, gradient, and radiofrequency (RF) fields used for MRI are thought to be detrimental to pacemaker function and possibly cause harm to patients undergoing MRI examinations. Potential adverse interactions between pacemakers and MRI include heating (3), induction of ventricular fibrillation (4), rapid atrial pacing (5), pacing at multiples of the RF pulse and associated rapid ventricular pacing (3,6–9), reed switch malfunction (6,10), asynchronous pacing (6,7,9), inhibition of pacing output (3,5,6,11), alteration of programming with potential damage to the pacemaker circuitry (5), or movement of the device (12–15). Notably, harmful effects to patients have also been documented. To date, 10 deaths have been attributed to MRI procedures in patients with pacemakers (16,17). However, these fatalities were poorly characterized, and no electrocardiographic (ECG) data were available (12). Importantly, no deaths have been reported during physician-supervised MRI procedures (12).
Numerous patients with implanted pacemakers have undergone MRI either inadvertently or during purposeful, monitored attempts to perform much needed examinations (11,18–26). These data suggest that MRI procedures are not as detrimental as once thought. Patients with pacemakers have been successfully imaged using magnetic resonance (MR) systems operating at static field strengths ranging from 0.35-Tesla (T) to 1.5-T without any clinically adverse events (18–25). The investigators of these studies and others have also suggested certain strategies for performing safe MRI procedures. These strategies included programming the pacemaker device subthreshold (13,18,19,21)or to asynchronous mode (3,9,11,13,22,27), programming to a bipolar lead configuration if possible (13,26), only imaging non–pacemaker-dependent patients (13,25), limiting exposure to RF power during MRI (9,13,19), and only performing MRI examinations if the pulse generator is positioned outside of the bore of the MR system (20). Some have even favored explanting the pulse generator prior to MRI (28).
Previous investigations have reported safety in dogs (29)and patients undergoing MRI procedures using MR systems operating at <1.5-T (10,11,19,20,23)and in a small number of patients at static magnetic field strengths at 1.5-T (11,18,21,22). Of note is that a recent study has identified safe criteria for a neurostimulation system (which is a pacing device for the brain with bilateral leads and dual implantable pulse generators) at 1.5-T (30).
Before the present investigation, no prospective study was performed in association with MR systems operating at 1.5-T. Therefore, this prospective study of pacemaker patients in need of MRI/magnetic resonance angiography (MRA) procedures was undertaken to establish a risk/benefit profile that might broaden access to this important diagnostic imaging modality.
Between December 22, 1999, and December 12, 2002, a total of 47 sequential patients with a permanent pacemaker and any clinical indication for MRI procedures were enrolled in a Hillcrest Medical Center (Tulsa, Oklahoma) Institutional Review Board-approved protocol and underwent 56 different MRI examinations at the Oklahoma Heart Institute MRI Center. Patients were counseled regarding the risks of the MRI, and informed consent was obtained in all patients. A total of seven pacemaker patients had undergone seven medically necessary MRI/MRA procedures, with informed consent, before the development of the Institutional Review Board protocol. Only pacemaker-dependent patients were excluded from this study, although one such patient inadvertently underwent MRI examination.
Data were obtained on all the patients, yielding a total of 54 sequential patients and 62 MRI/MRA examinations. An examination was considered separate if the patient exited the MR system room after the study was completed. Repeat examinations were done on eight patients, with one patient receiving three separate examinations. No limitations were placed on the type or duration of the MRI procedure, pacemaker, or lead models, nor proximity of the imaged anatomy relative to the pacemaker (Tables 1 and 2). ⇓⇓All MRI/MRA studies were performed using an MR system operating at a static magnetic field strength of 1.5-T (GE Signa CV/i, General Electric Medical Systems, Milwaukee, Wisconsin). This MR system has high performance gradients with a slew rate of 150 T/m/s, a rise time of 268 μs, and an amplitude of 40 mT/m.
An electrophysiologist interrogated each pacemaker. Data regarding initial programming, capture, and sensing thresholds, as well as lead impedances, were obtained. Rate histograms were cleared. Pacemakers were not programmed to the asynchronous mode or subthreshold. All patients were asked to report symptoms, especially sensations of heat or movement, in the area of the pacemaker. Artifacts were noted on the MR images, when present.
During MRI, continuous monitoring was performed of the ECG signal and of symptoms through voice contact with the patient via an intercom. Both an electrophysiologist and resuscitation equipment were available during all MRI examinations.
Symptoms and patient comments were recorded on a reporting form. Interrogations for programmed parameters, capture and sensing thresholds, lead impedances, battery voltage, and rate histograms were performed immediately upon exiting the MR system. Appropriate adjustments to the capture thresholds of the pacemakers were made when indicated.
All statistical analyses were conducted with SAS version 8.2 software (SAS Institute Inc., Cary, North Carolina). For each study subject, atrial and ventricular pacing thresholds were measured before and after each MRI examination. Two primary outcomes were evaluated: “any change” and “any significant change” in pacing threshold after MRI. The outcomes were parameterized as binary responses (Yes/No), and changes were evaluated for an association with chamber (atrial vs. ventricular), anatomy (above or below diaphragm was used for ease of statistical analysis), peak RF power, specific absorption rate (SAR), and time from lead implant to MRI examination. Peak SAR is a continuous variable, and its effect on pacing threshold change was evaluated in a logistic regression model. For all other analyses, a 2×2 contingency table is presented along with the corresponding p value. If an expected cell count for any 2×2 contingency table is <5, the reported p value is from Fisher exact test. Otherwise, reported p values are from the continuity-corrected chi-square test. In addition, multivariate logistic regression analysis was performed to evaluate each of the effects listed above in a pair-wise fashion, and possible interactions were explored. The primary analyses treated the repeated measurements for those eight patients with multiple MRI examinations, as well as atrial and ventricular leads in the same patient, as independent observations. However, a repeated measures analysis under a logistic model using generalized estimating equations methodology was also conducted to account for any within-patient correlation among repeated measurements.
After the MRI examinations, no episodes of loss of capture or changes in lead impedances or battery voltages were noted. In addition, no damage to pacemaker circuits or movement of the pulse generator was observed.
Patient symptoms were mild and infrequent. Two patients had reported symptoms. One patient felt his pacemaker vibrate during a coronary artery imaging sequence that used a spiral k-space acquisition, but this was transient. Another patient felt palpitations during a time of pacemaker inhibition. No MRI procedures were terminated as a result of any of the observed symptoms. Pacemakers also did not interfere with the interpretation of the MR images.
Of the 62 examinations, only 61 were evaluated, because one pacemaker was at the end-of-life. This yielded data on 107 leads, of which 48 were atrial and 59 were ventricular. Each lead was evaluated for “any change” in pacing threshold and “any significant change.” A significant change was defined as a change >1 voltage or pulse width increment or decrement, because changes smaller than that can be encountered frequently by chance.
Pacing thresholds changed in 40 of 107 (37%) leads. Of the 40 leads that changed, only 10 (9.4%) leads were judged to be a significant change. Only 2 of the 107 (1.9%) leads required a change in programmed output.
Changes were evaluated for an association with cardiac chamber, anatomy, peak SAR, and time from pacemaker lead implant to MRI examination. With regard to cardiac chamber analysis, 19 of the 48 atrial pacing thresholds (39.5%) changed after MRI, and 6 of these were significant (12.5%). Similarly, 21 of the 59 ventricular pacing thresholds changed after MRI (35.6%), 4 of which were significant (6.8%). There was no association between changes in pacing thresholds and cardiac chamber, with chi-square p values of 0.82 and 0.50 for “any change” and “any significant change,” respectively (Tables 3 and 4). ⇓⇓A repeated measures analysis was also conducted to account for the within-patient correlation between changes in ventricular and atrial pacing thresholds. This analysis demonstrated independence of atrial and ventricular pacing threshold changes within a patient (p = 0.6 for “any change” and p = 0.2 for “any significant change”). Therefore, all further analyses were conducted using each lead as the unit for analysis, and the atrial and ventricular leads were combined.
For statistical analysis, the imaged anatomical location was defined and grouped as above or below the diaphragm. One patient with both an atrial and ventricular lead had an MRI procedure both above and below the diaphragm during the same examination and, therefore, was excluded from the analysis. Twenty-six of 61 (42.6%) pacing thresholds changed after an MRI above the diaphragm, and 7 (11.5%) of these were significant. Fourteen of 44 (31.8%) pacing thresholds changed after an MRI below the diaphragm, 3 (6.8%) of which were significant. The chi-square p values associated with anatomy are 0.36 for “any change” in pacing threshold and 0.64 for “any significant change.” Therefore, no association existed between changes in pacing threshold and region of anatomy that underwent imaging (Tables 5 and 6). ⇓⇓
The log-odds ratio of any change in pacing threshold versus peak SAR was 0.41, with a corresponding p value of 0.14. As SAR increases by one unit, the odds of experiencing any change in pacing threshold increase 50%. However, the odds of a significant change in pacing threshold decrease as SAR increases (log-odds = −0.17; p = 0.74). Because the outcome is not statistically significant and the log-odds ratios are relatively close to 1, the disparate directional association is likely due to random variation.
Time from lead implant to the MRI examination was then examined to see whether this factor was related to threshold changes. Although over 40% of recent implants (0 to 4 years prior to MRI examination) experienced a change in pacing thresholds compared with approximately 20% for older leads, the comparison across implant times was not statistically significant for “any change” in pacing thresholds (p = 0.11; Fisher exact test) or for “any significant change” in thresholds (p = 0.63; Fisher exact test) (Tables 7 and 8). ⇓⇓
Results of the multivariate logistic regression analyses demonstrated no interactive effects, with p values for the interaction tests ranging from 0.37 (anatomical location by SAR) to 0.64 (anatomical location by chamber). Effects of each predictor in the multivariate analyses were consistent with that predictor's effect in the multivariate model.
Additionally, changes were seen on the ECG during testing that were clinically without symptoms or consequence. Atrial over-sensing of electromagnetic noise leading to ventricular tracking at the upper rate limit of the pacemaker was seen in two patients. Another episode of atrial over-sensing led to inhibition of pacing in an AAI system. Ventricular over-sensing with pacing inhibition for one heart beat occurred in one patient. Five patients exhibited transient non–magnet-mode behavior, most likely due to an open reed switch. Finally, one pacemaker with a minute ventilation feature sensed noise as an increase in minute ventilation and, therefore, paced at the upper rate limit. All other devices demonstrated magnet-mode behavior during the MRI procedure. No episodes of pacing above the upper rate limit were noted, and no arrhythmias were encountered.
Magnetic resonance imaging has become the ideal test in many areas of clinical medicine. Because of its high spatial and contrast resolution, it is now the primary diagnostic imaging test of choice for disorders of the central nervous and musculoskeletal systems. It is also important for oncological and certain cardiovascular disorders. In certain clinical instances, denying a patient an MRI procedure may have a significant effect on the patient's care. Subsequently, this could prove to have a major impact on overall public health. The most recent data from 2002 suggest that 2.4 million people in the U.S. now have permanent implanted pacemakers. An additional 430,000 individuals are estimated to undergo pacemaker implantation in 2003 (Scot Milchman, Guidant Corporation, St. Paul, Minnesota, personal communication, April 28, 2003). This large group of patients is presently being denied a potentially important diagnostic examination.
In 1999, a Japanese study polled 1,567 pacemaker patients and reported that 17% needed MRI procedures during the previous year but were denied because of the presence of an implanted cardiac pacemaker (31). Applying this percentage to the current number of people in the U.S. with pacemakers yields 408,000 individuals. This is a major public health issue, as these patients may have to undergo an inferior imaging procedure or an invasive examination, leading to a possible missed diagnosis and/or increased morbidity.
Findings from the present study
In the current study, to examine risk in the broadest possible population, no restrictions were placed on the anatomy that underwent imaging, the type of pulse sequence, or the conditions used for MRI, nor on the type of pacemaker present in the patient. Only pacemaker-dependent patients were excluded (with the exception of one patient who inadvertently underwent MRI) so as to eliminate problems if pacing was inhibited during the MRI procedure. Furthermore, the decision was made not to program to an asynchronous mode because pacemakers automatically enter the asynchronous mode when in the presence of a powerful static magnetic field. Pacemakers were also not programmed to subthreshold because it was believed that the risk of developing ventricular fibrillation during asynchronous pacing was extremely low. This assumption is supported by the fact that numerous times transtelephonic pacemaker interrogations occur without incident on a daily basis. However, resuscitation equipment and advanced cardiac life support (ACLS)-trained personnel were readily available if problems arose for those patients in the MR environment.
The threshold data were analyzed with respect to SAR, anatomy imaged, cardiac chamber, and time from lead implant to the MRI examination. Any change in threshold was considered important, even though only increases in these values are detrimental.
One study suggests that limiting the distance from the pacemaker pulse generator to the anatomical region being imaged is beneficial to MRI safety (20). Therefore, the data were analyzed with respect to the anatomy that underwent imaging. For ease of statistical analysis, the MRI examinations were grouped into above and below the diaphragm. Neither anatomical region was associated with pacing threshold changes.
Pacing thresholds were also examined as a function of cardiac chamber. It was theorized that the atrium, because of its smaller mass with possible less heat transfer, would more readily undergo threshold changes than would the ventricle. However, the atrium was no more likely to develop pacing threshold changes than was the ventricle.
Because the level of the SAR (i.e., the dosimetric term used to indicate the amount of RF energy used during an MRI procedure) has been shown to be responsible for the relative amount of heating that occurs for medical implants, including pacemakers and neurostimulation systems (2,3,31,32), it was believed that the higher the SAR the more likely the pacemaker lead would be to heat and cause subsequent threshold changes. However, there were no findings to support this suspicion. Threshold changes were unrelated to the peak whole-body–averaged SARs associated with the MRI procedures for the patients in the present study.
We theorize that pacing threshold changes may occur because of electrode heating. Pacing leads have been shown to heat in vitro to 88.8°C (3). This is well above ablation temperature, and if it occurred in vivo the pacemaker device would no longer be able to pace the myocardium. Therefore, some intrinsic heat transfer must be occurring and is most likely due to flowing blood and myocardial mass. We postulate that the declines of threshold represent mild levels of heating at the lead-tissue interface, whereas temperature rises are secondary to higher levels of thermal injury. Proof of this theory is absent but is supported by observations that mirror these changes in the electrophysiology laboratory. Despite the pacing threshold changes observed in the present study, no serious adverse events were observed despite the use of the 1.5-T MR system and the lack of exclusions for MRI techniques and body parts that underwent examinations.
The significant pacing threshold changes noted during our study were infrequent and all easily addressed with subtle programming changes. The energy increases that were needed to accommodate the rise in thresholds were minor and did not impair the safe performance of the pulse generators. Despite the labeling of these changes as significant, they were of no clinical consequence. Changes of these magnitudes are commonly observed in the daily practice of cardiology.
Finally, minor ECG changes were encountered, and these were varied and transient as described previously. Pacing inhibition was seen briefly in one atrial and one ventricular lead. It was because of this potential problem with inhibition, especially in the ventricular leads, that we chose not to include pacemaker-dependent patients in the study. The pacemaker inhibition phenomenon has been reported previously (5,11). Depending on the length of time of the inhibition, there could be a prolonged period of asystole. For this reason or because of induction of ventricular fibrillation, we postulate that some pacemaker patients have previously died in association with an MRI procedure (16,17). Non–magnet-mode response occurred in five patients, suggesting that the reed switch can open and close during MRI. This is consistent with previous reports (14,33)and was not an issue clinically.
A minute-ventilation pacemaker sensed noise as an increase in minute ventilation and paced at the upper rate limit. This patient happened to undergo two MRI examinations, and during the second examination the minute-ventilation feature was disabled and no upper rate limit pacing was seen.
All of the above noted instances were brief, and none was associated with clinical consequences. Finally, electromagnetically induced noise was encountered occasionally on telemetry. This was monitored closely, because it can resemble serious cardiac arrhythmias.
Patient symptoms were mild and transient. One patient reported vibration, but this could not be visually confirmed. Vibration of a device in association with an MRI procedure has been previously reported (34). One other patient had palpitations. This coincided with the inhibition of pacing in the ventricular lead. None of these symptoms required cessation of MRI. The presence of the pacemakers also did not affect the operation of the MR system or the interpretation of the MR images.
Other electronically activated devices
In the past, the presence of an electronically activated implant was considered a strict contraindication for a patient or individual in the MR environment (2). However, over the years, various studies have been performed to define safety criteria for electronic devices including pacemakers, implantable cardioverter defibrillators, neurostimulation systems, cochlear implants, a drug infusion pump, and a bone fusion stimulator (19,20,31,32,34–43). As such, findings have indicated that, if highly specific guidelines are followed, MRI procedures may be conducted safely in patients with electronically activated implants (19,20,31,32,34–49). Notably, some of these electronically activated devices have received approval of “MR safe” labeling claims from the U.S. Food and Drug Administration. Accordingly, it is hoped that cardiac pacemaker manufacturers will be encouraged by the results of the present study to pro-actively support and/or conduct investigations directed toward identifying MRI safety criteria for their respective devices. This will ultimately have a substantial impact on patient management and the overall health care of pacemaker patients who might require MRI procedures.
Given the infinite possibilities of pacing systems and cardiac and lead geometry, as well as variable static, gradient, and RF electromagnetic fields and conditions used for MRI procedures, the absolute safety of pacemaker and MRI interactions presently cannot be assured. However, given appropriate patient selection as well as continuous monitoring and preparedness for resuscitation efforts with ACLS-trained personnel in attendance, performance of MRI procedures on non–pacemaker-dependent patients may be achieved with reasonable safety, even at static magnetic field strengths of 1.5-T with an acceptable risk/benefit profile.
Finally, it should be noted that the findings described herein are highly specific to the MR system, the software version running the scanner, MRI conditions, and types of pacemakers and lead systems present in the patients. Other factors that potentially impact MRI safety for patients with pacemakers will need to be defined.
Given the results of the current study, the following guidelines should be followed to reduce the chance of an adverse pacemaker/MRI interaction: 1) obtain informed consent; the patient needs to understand that there is still a potential risk; 2) have emergency equipment and ACLS-trained personnel readily available; 3) interrogate the pulse generator immediately before and after MRI; 4) disable the minute ventilation feature; 5) maintain voice contact throughout the procedure and continuously monitor heart rhythm and rate; 6) because subtle pacemaker programming changes occasionally are needed, a physician facile in the ways of pacemaker programming needs to be available; and 7) with respect to subthreshold output programming, previous investigators have suggested this approach to minimize induction of ventricular fibrillation. This approach is reasonable but may not be necessary when the aforementioned guidelines are followed.
Possible study limitations
A possible limitation of this study is that the effects of MRI-related heating were not directly measured. Also, no comment could be made regarding pacemaker-dependent patients, as they were intentionally not included in the study population.
In this series of 62 non–pacemaker-dependent patient examinations, performance of unrestricted MRI procedures using a 1.5-T MR system was found to have an acceptable safety profile. Patient symptoms were mild and transient and did not lead to discontinuation of the examinations. Significant alteration of the pacing threshold was found in a small number of leads tested. These threshold changes required a programmed output change in only two leads and were of no clinical consequence. The threshold changes were unrelated to cardiac chamber, anatomical location, and time from lead implant to MRI examination and highest whole body SAR. Therefore, the belief in the presence of a pacemaker as an absolute contraindication to MRI should be re-evaluated.
- advanced cardiac life support
- magnetic resonance
- magnetic resonance angiography
- magnetic resonance imaging
- specific absorption rate
- American College of Cardiology Foundation
- Shellock F.G
- Hayes D.L,
- Holmes D.R,
- Gray J.E
- ↵Gimbel JR. Implantable pacemaker and defibrillator safety in the MR environment: new thoughts for the new millennium. In: 2001 Syllabus, Special Cross-Specialty Categorical Course in Diagnostic Radiology: Practical MR Safety Considerations for Physicians, Physicists, and Technologists. Oak Lawn, IL: Radiological Society of North America, 2001;69–76
- Duru F,
- Luechinger R,
- Scheidegger M.B,
- Luscher T.F,
- Boesiger P,
- Candinas R
- Food and Drug Administration. Center for Devices and Radiological Health MR Product Reporting Program. Available at: http://www.fda.gov/cdrh/mdr/. Accessed May 14, 2003
- Inbar S,
- Larson J,
- Burt T,
- Mafee M,
- Ezri M
- Iberer F,
- Justich E,
- Stenzl W,
- Tscheliessnig K.H,
- Kapeller J
- Coman J.A,
- Martin E.T,
- Ramza B.M,
- Margolis P.D,
- Fair R.D
- Juratli N,
- Sparker J,
- Gimbel J.R,
- Wilkoff B
- Finelli D.A,
- Rezai A.R,
- Ruggieri P,
- et al.
- Luechinger R,
- Duru F,
- Zeijlemaker V,
- Scheidegger M,
- Boesiger P,
- Cardinas R
- Reilly P,
- Diamant A.M