Conventional pacemaker and implantable cardioverter-defibrillator product labeling currently cautions against exposure to magnetic resonance imaging (MRI). However, there is a growing clinical need for MRI, without an acceptable alternative imaging modality in many patients with cardiac devices. The purpose of this study was to determine the risk of MRI at 1.5 T for patients with cardiac devices by measuring the frequency of device failures and clinically relevant device parameter changes. Data from a single-center retrospective review of 109 patients with pacemakers and implantable cardioverter-defibrillators (the MRI group) who underwent 125 clinically indicated MRI studies were compared to data from a prospective cohort of 50 patients with cardiac devices who did not undergo MRI (the control group). In the MRI group, there were no deaths, device failures requiring generator or lead replacement, induced arrhythmias, losses of capture, or electrical reset episodes. Decreases in battery voltage of ≥0.04 V occurred in 4%, pacing threshold increases of ≥0.5 V in 3%, and pacing lead impedance changes of ≥50 Ω in 6%. Although there were statistically significant differences between the MRI and control groups for the mean change in pacing lead impedance (−6.2 ± 23.9 vs 3.0 ± 22.1 Ω) and left ventricular pacing threshold (−0.1 ± 0.3 vs 0.1 ± 0.2 V), these differences were not clinically important. In conclusion, MRI in patients with cardiac devices resulted in no device or lead failures. A small number of clinically relevant changes in device parameter measurements were noted. However, these changes were similar to those in a control group of patients who did not undergo MRI.
Magnetic resonance imaging (MRI) has been considered a relative contraindication for patients with non-MRI-compatible pacemakers and implantable cardioverter-defibrillators (ICDs). A scientific statement by the American Heart Association published in 2007 provided guidelines for device settings and safety monitoring during MRI but did not endorse the performance of MRI for patients with pacemakers and ICDs. The purpose of the present study was to evaluate the risk of MRI at 1.5 T in a retrospective group of patients with implanted cardiac devices who underwent clinically indicated MRI, by determining the frequency of adverse clinical events and clinically relevant device parameter changes. Because there is a lack of published data on the variability of parameter measurements that normally occur in cardiac devices, parameter measurement data in a separate group of patients with pacemakers and ICDs who did not undergo MRI (the control group) were prospectively collected to define the magnitude of device background parameter variability.
Methods
Medical records were reviewed for all patients with permanent pacemakers and ICDs who underwent clinically necessary MRIs from February 2006 to March 2009 at a single institution (the MRI group). MRI was ordered by the patient’s treating physician, who determined that there was no acceptable alternative imaging study and that the potential benefit of the diagnostic data obtained by MRI significantly outweighed the potential risk for device failure. Patients who were pacemaker dependent (intrinsic ventricular rate <40 beats/min or symptomatic bradycardia with a heart rate ≥40 beats/min) and those who were not pacemaker dependent were included in the protocol and subsequent analysis. Participants in the control group were recruited during standard device interrogation clinic appointments from August 2008 to June 2009 and underwent 2 device interrogations approximately 1 hour apart without an intervening MRI study (control group). The study was approved by the institutional review board; patients in the control group gave written consent.
In the MRI group, studies were performed using a Siemens Magnetom Symphony 1.5 T scanner (Siemens Healthcare, Erlangen, Germany) using standard imaging protocols for the evaluation of the prescribed anatomic region. There were no prespecified specific-absorption-rate limits established for cardiac device patients. Data relating to specific-absorption-rate limits and values were not recorded during routine clinical imaging and thus were not available for analysis.
Device interrogation was performed immediately preceding the MRI study. Parameters interrogated and recorded were battery voltage, pacing thresholds, P-wave and R-wave amplitudes, pacing lead impedances, and high-voltage lead impedance (only when obtained by painless interrogation). Thresholds were preferably obtained after disabling automatic capture measurement algorithms, using a bipolar lead configuration at a pulse width of 0.4 ms. In pacemaker-dependent patients, the pacemaker was reprogrammed to an asynchronous pacing mode, and the magnet response was disabled when possible. For patients who were not pacemaker dependent, pacing and sensing functions were deactivated. For patients with ICDs, tachyarrhythmia therapies were disabled. Patients were monitored throughout the procedure with continuous cardiac rhythm recording and pulse oximetry. A cardiologist with experience in cardiac device programming who was able to place and use a temporary external cardiac pacemaker was present throughout the MRI study.
Immediately after the MRI study, a repeat interrogation was performed using a protocol identical to the prescan interrogation, and prescan device parameters were restored. Threshold measurements were performed at the same pacing rate used for the prescan device interrogation. Reprogramming was performed at the discretion of the supervising cardiologist to maintain appropriate threshold output and sensitivity safety margins. Interrogations in the control group were performed similarly.
The primary study end points in the MRI group were: death during the MRI study, device or lead failure requiring immediate replacement, induced atrial or ventricular arrhythmias during the MRI study, loss of pacemaker capture, and electrical reset of the device (to default parameter settings). For secondary end points, cut-off values were selected on the basis of the potential need for either device reprogramming or increased surveillance appropriate in normal clinical practice: (1) battery voltage decrease ≥0.04 V, (2) pacing lead threshold increase ≥0.5 V at a fixed pulse width duration of 0.4 ms (leads that were measured at a pulse width duration other than 0.4 ms were excluded from analysis), (3) P-wave amplitude decrease ≥50%, (4) R-wave amplitude decrease ≥25%, (5) absolute pacing lead impedance change ≥50 Ω, and (6) absolute high-voltage lead impedance change ≥3 Ω.
To compare the MRI and control groups with respect to parameter changes, and to account for the fact that some patients had >1 MRI study, linear mixed-model analysis was used to compare the MRI and control groups with respect to battery voltage change, P- and R-wave percentage changes, and high-voltage impedance change, while adjusting for type of device (pacemaker or ICD) and pacemaker dependency (yes or no). The same mixed-model approach was used for pacing lead threshold and impedance change, except that the model also took into account lead measurements from multiple chambers within the same patient. SPSS version 14.0 (SPSS, Inc., Chicago, Illinois) was used for linear mixed-model analyses. All tests were 2 tailed, and p values <0.05 for continuous variables were considered statistically significant. The sample size in the MRI group was the total number of patients with either pacemakers or ICDs who underwent clinically indicated MRI at our institution from February 2006 to March 2009. The sample size in the control group was determined by the number of eligible patients who agreed to participate within the specified time frame.
Results
In the MRI group, 109 patients with pacemakers and ICDs underwent 125 consecutive, clinically indicated MRI studies. A single scan was performed in 95 patients, 12 underwent 2 separate studies on different days, and 2 patients underwent 3 studies on different days. Three patients entered the scanner and were exposed to the static magnet field but did not undergo successful imaging (claustrophobia in 1 patient, unacceptable artifact from a biventricular ICD in 1 patient, and a pacemaker magnet response in 1 patient who underwent successful imaging with the magnet response disabled on a subsequent day). The clinical and parameter change data from these 3 patients were included in the MRI group analysis without qualification or distinction. Baseline characteristics of the MRI and control groups are listed in Table 1 .
| Variable | MRI Group | Control Group |
|---|---|---|
| (n = 125 Studies in 109 Patients) | (n = 50 Patients) | |
| Age (years) | 74 ± 11 | 75 ± 10 |
| Men | 76 (61%) | 32 (64%) |
| Years since implantation ⁎ | 2.6 ± 2.2 | 2.4 ± 2.3 |
| Minutes between interrogations | 49 ± 21 | 55 ± 14 |
| Pacer dependent † | 29 (27%) | 8 (16%) |
| Devices | 125 | 50 |
| Pacemaker, single chamber | 3 (4%) | 2 (4%) |
| Pacemaker, dual chamber | 79 (63%) | 20 (40%) |
| Pacemaker, biventricular | 3 (2%) | 0 |
| ICD, single-chamber | 4 (3%) | 5 (10%) |
| ICD, dual-chamber | 20 (16%) | 15 (30%) |
| ICD, biventricular | 16 (13%) | 8 (16%) |
| Leads | 259 | 101 |
| Right atrial | 115 (44%) | 43 (43%) |
| Right ventricular | 125 (48%) | 50 (50%) |
| Left ventricular | 19 (7%) | 8 (8%) |
| Anatomic region scanned | 157 | Not applicable |
| Brain | 49 (31%) | |
| Spine, lumbar | 28 (18%) | |
| Spine, cervical | 24 (15%) | |
| Spine, thoracic | 15 (10%) | |
| Cardiac | 7 (4%) | |
| Knee | 7 (4%) | |
| Pelvis | 7 (4%) | |
| Abdomen | 4 (3%) | |
| Orbit | 4 (3%) | |
| Shoulder | 4 (3%) | |
| Other | 8 (5%) |
⁎ Mean time between generator implantation and interrogation, independent of age of lead.
† Pacer dependence data were not available for 19 of 125 MRI studies.
In the MRI group, there were no deaths, acute device or lead failures requiring replacement, new atrial or ventricular arrhythmias, losses of pacing capture, or electrical device resets noted during the MRI studies ( Table 2 ). The overall likelihood of having any parameter change event was 18% (22 of 125) in the MRI group and 22% (11 of 50) in the control group, which did not undergo MRI. The frequency of parameter change events in the MRI group is reported in Table 2 . Applying the same parameter-change-event cut-off values used for the MRI group to the control group, 0% of the control group (0 of 44) had battery voltage change events, 1% (1 of 90) had a pacing lead threshold change event, 3% (1 of 34) had a P-wave amplitude change event, 5% (2 of 40) had R-wave change events, 4% (4 of 94) had pacing lead impedance change events, and 9% (2 of 23) had high-voltage impedance change events. A slightly higher proportion of battery voltage, pacing lead threshold, pacing lead impedance, and high-voltage impedance parameter change events were observed in the MRI group, and a slightly higher proportion of P- and R-wave amplitude events were observed in the control group.
| End Point | Number of Events/Number of Events Possible (%) [95% Confidence Interval] |
|---|---|
| Primary end point events | |
| Death | 0/125 (0%) [0%–3%] |
| Device failure | 0/125 (0%) [0%–3%] |
| Lead failure | 0/259 (0%) [0%–1%] |
| Atrial/ventricular arrhythmia | 0/125 (0%) [0%–3%] |
| Loss of capture | 0/125 (0%) [0%–3%] |
| Electrical reset | 0/125 (0%) [0%–3%] |
| Secondary end point events | |
| Battery voltage decrease ≥0.04 V | 4/90 (4%) [2%–11%] |
| Pacing lead threshold increase ≥0.5 V | 5/195 (3%) [1%–6%] |
| P-wave amplitude decrease ≥50% | 1/65 (2%) [0.3%–8%] |
| R-wave amplitude decrease ≥25% | 2/74 (3%) [1%–9%] |
| Pacing lead impedance change ≥50 Ω | 13/210 (6%) [4%–10%] |
| High-voltage impedance change ≥3 Ω | 2/17 (12%) [3%–34%] |
The distributions of the parameter-change data in the MRI and control groups are presented in Figure 1 . The mean change in pacing lead impedance in the MRI and control groups was statistically different but did not reach a threshold of clinical importance ( Table 3 ). There were no significant differences between the MRI and control groups with respect to battery voltage, P- and R-wave amplitudes, and high-voltage lead impedance. However, linear mixed-model analysis for pacing lead threshold yielded a significant interaction effect between the chamber in which leads were located (right atrium, right ventricle, or left ventricle) and the group (MRI and control), which precluded a main-effects comparison. A small mean decrease in left ventricular pacing threshold in the MRI group and a small mean increase in the control group were noted.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
