MISCELLANEOUS TOPICS
Effect of MagneticResonanceImaging on DDDPacemakers JAY A. ERLEBACHER,
MD, PATRICK T. CAHILL, PhD, FRANK PANNIZZO, and R. JAMES R. KNOWLES, PhD
A previous study suggested the safety of exposing patients with certain pacemakers models to magnetic resonance imaging (MRI). However, the function of a variety of more advanced DDD pacemakers and the effect of higher magnetic and radiofrequency (rf) field strengths has not been reported. In the present study 4 different DDD pacemakers (Cordis 233F, Intermedic 283-01, Medtronic 7000A, and Pacesetter 283) were tested in a saline phantom under several conditions and with various imaging sequences. Pacemaker output was monitored using electrocardiographic telemetry. All units paced normally in the static magnetic field. However, during imaging, all units malfunctioned, with total inhibition of atrial and ventricular output in
MS,
3 of the pacemakers. In the fourth pacemaker, ventricular backup pacing was activated at high rf pulse repetition rates. However, the MRI scanner could trigger atrial output in this pacemaker at rates of up to 800/minute. All malfunctions were a result of rf interference, whereas gradient and static magnetic fields had no effect. Thus, despite magnetic field strengths adequate to close pacemaker reed switches, rf interference during MRI may cause total inhibition of atrial and ventricular output in DDD pacemakers, and can also cause dangerous atrial pacing at high rates. MRI should be avoided in patients with these DDD pacemakers. (Am J Cardiol 1986;57:437-440)
M
agnetic resonance imaging (MRI) is becoming recognized as a clinically useful and safe tool for the diagnosis of a variety of disease states. Despite the apparent lack of biologic effects from MRI, the safety of performing MRI in patients with implanted electrical devices such as pacemakers has not been convincingly demonstrated. MRI requires 3 electromagnetic fields: a strong static magnetic field to orient the hydrogen nuclei, a pulsed gradient magnetic field to select a particular imaging volume and a pulsed radiofrequency (rf] elec-
trical field to perturb the hydrogen nuclei. These electromagnetic fields may cause damage to the magnetic reed switch in the pacemaker, alter programmed variables and cause unexpected changes in pacing function. Fetter et all reported that selected pacemakers of a single manufacturer may operate safely during MRI scanning. However, several pacemakers in their study did malfunction, and the conclusions regarding safe operation could only be made regarding the l,5OOgauss, 1,000-W MRI instrument that they tested. They suggested that dual-chamber pacemakers may be particularly susceptible to electromagnetic interference from MRI scanning. We evaluated the function of DDD pacemakers made by 4 different manufacturers. These units were tested in vitro in a 5,000-gauss,10,000-W MRI scanner, which is one of the more powerful scanners available. MRI systems with even stronger field strengths are now being evaluated.
From the Departments of Medicine and Radiology, Cornell University Medical College, New York, New York, and the Montefiore Hospital and Medical Center Cardiac Pacemaker Center, Bronx, New York. Manuscript received June 24, 1985; revised manuscript received July 26,1985, accepted July 30,1985. Address for reprints: Jay A. Erlebacher, MD, Cardiology Division, The New York Hospital-Cornell Medical Center, 525 East 68th Street, New York, New York 10021. 437
438
DDD PACEMAKER
MALFUNCTION
WITH MAGNETIC
RESONANCE
Methods Magnetic resonance imaging system: The MRI scanner used in these studies is a Technicare (Johnson & Johnson) Teslacon. The superconducting magnet operates at a static magnetic field strength of 5,000 gauss. During imaging, this field is altered slightly by rapidly switching magnetic gradients. The maximal change in magnetic field strength is 0.25 gauss/cm, with a rise time of approximately 2 ms. Thus, the maximal rate of change is approximately 120 gauss/cm/s. During imaging, rf pulses are delivered with a lokW instantaneous peak power at a frequency of 20.91 MHz. The rf power is delivered in the form of 90degree and 180-degree gaussian pulses with a width of 2 to 8 ms. The go-degree pulse is followed by one or more 180-degree pulses at intervals that are odd multiples of TE1/2, where TE1 = echo time. TE1 ranges from 30 to 40 ms. The whole sequence is repeated at intervals known as the repetition time (TR), which generally varies from 2,500 ms to less than 100 ms in typical imaging sequences.
IMAGING
Pacemakers: Pacemakers made by 4 manufacturers were evaluated. These were the Cordis 233F, Intermedic 283-01, Medtronic 7000A and Pacesetter 283. Each unit was programmed to the DDD mode. Atria1 and ventricular sensitivity and output were programmed to nominal values. Electrical testing: Each pacemaker was tested for stability of programmed values before and after exposure to the MRI scanner. Pulse width and amplitudes, lower rate and, where possible, sensitivity were tested for stability before and again after MRI exposure. With 1 exception, each manufacturer’s pacemaker was tested with a pacemaker system analyzer manufactured by the pacemaker company. Because no pacemaker analyzer is made by Pacesetter, a Medtronic pacemaker analyzer was used to test the Pacesetter pacemaker. To determine whether the extraordinary magnetic field strengths encountered during MRI scanning permanently affect the pacemaker magnetic reed switch, the distance at which the standard pacemaker magnet provided by each manufacturer activated each unit’s reed switch was measured before and after exposure to the MRI scanner. Pacemaker monitoring during magnetic resonance imaging: Pacemakers and their associated leads were placed in a polyethelene container 33 cm wide, 23 cm long and 10 cm deep that was filled with saline solution. The pacemaker leads were held by plastic fixation devices in a shape typical of implanted leads. A Hewlett-Packard electrocardiographic (ECG) telemetry transmitter (model #78100A) was used to transmit ECG signals detected from standard patient ECG wires placed in the container filled with saline solution. A Hewlett-Packard telemetry receiver (model #78101A) was used to monitor and record pacemaker function on standard ECG paper at a chart speed of 25 mm/s. Pacemaker function was evaluated under 4 conditions: 12 feet from the MRI scanner, inside the MRI scanner with the static magnetic field on, inside the MRI scanner with static and gradient magnetic fields on, and inside the MRI scanner with the static magnetic field and rf field on. Usually, testing was done in the single slice mode with a TE1 of 30 ms and a TR of 500 ms. In selected studies, TRs as short as 75 ms were used.
Results
FIGURE 1. Demonstration of malfunction of 283-01 pacemaker. Top, normal atrioventricular (AV) sequential operation 12 feet from magnetic resonance imaging (MN) scanner. Top middle, normal AV sequential operation inside MRI scanner with gradient and radiofrequency (RF) fields off. Bottom middle, normal AV sequential operation inside MRI scanner with gradient fields on (repetition time = 500 ms, echo time = 30 ms). Bottom, inhibition of pacemaker output inside MRI scanner with radiofrequency field on. a = atrial output; A = MRI field artifact; SS = single slice imaging technique; v = ventricular output.
All 4 pacemakers produced atrioventricular pacing artifacts 12 feet from the superconducting magnet. All devices continued to pace normally inside the MRI magnet core in the absence of gradient magnetic and rf fields. When gradient magnetic fields were superimposed on the static field there was no change in pacing. However, when the gradient field was turned off and the rf field was applied, all 4 pacemakers had serious malfunctions. The 283, 283-01 and the 7000A pacemakers produced no visible output on the electrocardiogram when exposed to the IO-kW, 20.91-MHz rf pulses produced by the MRI system (Fig. I]. This abnormality was unexpected because the static magnetic field is at
February 15, 1986
least 100 times stronger than a typical pacemaker magnet. Applying a pacemaker magnet to a pacemaker ordinarily produces asynchronous DO0 operation even in the face of electromagnetic interference. The total inhibition of pacemaker output by the rf pulses during MRI scanning was therefore a surprising finding that has serious implications for pacemaker-dependent patients. In contrast to the total inhibition of output by rf in the above pacemakers, the 233F pacemaker produced atria1 output that appeared to be triggered by the onset of each rf pulse train (Fig. 2). The atria1 pacemaker rate could be driven by rf pulsing at all pulse train TRs from 600 to 75 ms [Fig. 3). ECG artifacts were shown to be produced by the pacemaker atria1 channel by separately testing pacemaker leads alone, the pacemaker generator alone, the generator f ventricular lead and the generator + atria1 lead. As the TR interval was shortened from 400 to 300 ms, ventricular output appeared at a rate of 55/minute and the pacemaker was found to be in the back-up mode, Thus, the 233F pacemaker was the only unit tested that continued to pace during MRI scanning.
FIGURE 2. Malfunction of a 233F pacemaker. Top,normal atrioventricular (AV) sequential operation 12 feet from magnetic resonance (MRI) scanner. Top middle, normal AV sequential operation inside MRI scanner with gradient and rf fields off. Bottom middle, normal AV sequential operation inside MRI scanner with gradient fields on. Boffom, atrial output triggered by MRI radiofrequency (RF) field at 500-ms intervals. Abbreviations as in Figure 1.
THE AMERICAN
JOURNAL.
OF CARDIOLOGY
Volume 57
4.38
However, no ventricular output was present until short TR intervals activated ventricular backup pacing. Furthermore, although atria1 output was always present during scanning, atria1 rates could be excessive, depending on the TR interval. None of the pacemakers changed their programmed or measured variables as a result of MRI scanning. With 1 exception, no changes occurred in the maximal magnet distance required to close the pacemaker reed switch and initiate DO0 pacing before and after MRI scanning. The 283 pacemaker reed switch could be activated 6.5 cm from the pacemaker magnet after MRI scanning vs 4.8 cm before MRI exposure.
Discussion Preliminary investigations by Fetter et all using a single manufacturer’s pacemakers have suggested that under some conditions, MRT scanning may be safe, given the understanding that asynchronous pacemaker operation will occur. However, Fetter et al noted malfuction of 2 of the 4 implantable pacemakers that were tested. We tested DDD pacemakers made by 4 different manufacturers inside an MRI scanner with a magnet more than 3 times more powerful and a rf
FIGURE 3. Atrial output tracking at various magnetic resonance imaging radiofrequency (RF) repetition intervals (TR). In the 3 panels atrial tracking (a) is seen at 600-, 400- and lOO-ms intervals. Boffom, ventricular backup pacing (V,,) has been actuated at a rate of 55 beats/min. SS = single slice imaging technique.
440
DDD PACEMAKER
MALFUNCTION
WITH MAGNETIC
RESONANCE
source 10 times more powerful than that used by Fetter et al. We found that all 4 pacemakers tested had serious. malfunctions during MRI scanning. Three pacemakers had no output during scanning, implying that patients with such pacemakers may have asystole when scanned under these conditions. One of the pacemakers with such malfunction was the same model as that appeared to operate normally when tested in the smaller MRI scanner used by Fetter et al. A fourth manufacturer’s unit produced atria1 output at the same rate as the MRI scanner’s rf TR. Unlike the above pacemakers with no output, patients with intact atrioventricular conduction would have ventricular contractions during MRI scanning. However, the pacemaker atria1 channel could “track” at rates up to 800 beats/min, placing patients with coronary artery disease at risk and possibly inducing atria1 arrhythmias in others. Patients with atrioventricular block would have no ventricular support unless ventricular backup pacing in this particular unit was deliberately or automatically activated. This did not occur automatically unless the TR interval was 300 ms or less. The cause of pacemaker malfunction during MRI was found to be the high-power pulsed rf field actuated during imaging. The static magnetic field was not harmful to pacemaker operation aside from the expected activation of asynchronous pacing. Likewise, the gradient magnetic fields did not interfere with pacing. Programmed settings were not altered, and only small changes in magnetic reed switch operation were noted.
IMAGING
Given these findings, we believe that MRI scanning should be avoided in patients with implanted DDD pacemakers. Non-pacemaker-dependent patients with pacemakers that inhibit during MRI could be imaged with simultaneous ECG monitoring or pulse palpation. Patients with a pacemaker triggered during MRI scanning may be imaged with ventricular backup pacing activated. High rf pulse repetition rates should he avoided because of the associated rapid atria1 triggering. The likelihood of competitive atria1 and ventricular rhythms should also be considered because pacemaker operation is asynchronous. The MRI scanner and pacemaker specificty noted ~ in this and the previbusly published study1 suggests that caution be used before extrapolating the results of these studies to other pacemaker models and MRI scanners. In fact, small changes in the electrical circuitry within pacemaker models may change responses even when tested under otherwise identical conditions. The results of this and other studies should provide an impetus to pacemaker manufacturers to conduct investigations aimed at making pacemakers impervious to rf interference produced by MRI scanners. Physicians and those involved with MRI scanning should be made aware of the hazards associated with exposure of patients with pacemakers to MRI.
References 1. Fetter J, Aram G, Holmes DR, Gray J, Hayes DL. The effects of nuclear
magnetic resonance imogers on external and implontoble pulse generators. PACE 1984;7:720-727.