Author + information
- Received August 22, 2004
- Accepted October 18, 2004
- Published online March 15, 2005.
- Alexandre Trigano, MD*,* (, )
- Olivier Blandeau, BS*,
- Martine Souques, MD†,
- Jean Pierre Gernez, BS‡ and
- Isabelle Magne, PhD‡
- ↵*Reprint requests and correspondence:
Dr. Alexandre Trigano, Centre Hospitalier Universitaire Nord, Hôpital Nord, 13915 cedex 20, Marseille, France
Objectives This study examined the risk of interference by high magnetic flux density with permanent pacemakers.
Background Several forms of electromagnetic energy may interfere with the functions of implanted pacemakers. No clinical study has reported specific and relevant information pertaining to magnetic fields near power lines or electrical appliances.
Methods A total of 250 consecutive tests were performed in 245 recipients of permanent pacemakers during 12-lead electrocardiographic monitoring. A dedicated exposure system generated a 50-Hz frequency and maximum 100-μT flux density, while the electrical field was kept at values on the order of 0.10 V/m.
Results A switch to the asynchronous mode was recorded in three patients with devices programmed in the unipolar sensing configuration. A sustained mode switch was followed by symptomatic pacing inhibition in one patient. No effect on devices programmed in bipolar sensing was observed, except for a single interaction with a specific capture monitoring algorithm.
Conclusions The overall incidence of interaction by a magnetic field was low in patients tested with a wide variety of conventionally programmed pacemaker models. A magnetic field pulsed at power frequency can cause a mode switch and pacing inhibition in patients with devices programmed in the unipolar sensing configuration. The risk of interference appears negligible in patients with bipolar sensing programming.
Electromagnetic interference (EMI) with implanted pacemakers has been studied in vitro and in several clinical studies or reported from anecdotal daily life observations. Electromagnetic interference may be observed near high-voltage power lines and plants, transformers, or other structures or may be caused by electrical appliances held close to the chest. Although interference by strong electrical fields has been widely reported, EMI from magnetic fields has not been studied as intensively. Strong magnetic fields are present in industrial or occupational environments and emitted in day-to-day life by household appliances and some electronic surveillance articles. The safe limits of exposure to magnetic flux in recipients of implanted pacemakers remain to be established. Although simulations using a model of the human body have been presented, a single, nondefinitive clinical study has been published on this subject (1–3).
This study was designed to examine, in a large patient population, the behavior of implanted cardiac pacemakers in the presence of magnetic fields at power line frequency and 100-μT flux density, the value retained at 50 Hz in the European recommendations for general public exposure (1999/519/EC) (4). The objectives were to provide clinical data to international organizations responsible for establishing specific limits of exposure for recipients of permanent pacemakers.
The study design was approved by the Ethical Committee for Human Research of La Pitié-Salpétrière Hospital, University of Paris, France. All patients between 18 and 85 years of age presenting for routine ambulatory pacemaker follow-up during the study period were invited to participate. Written, informed consent was obtained from all patients. Pretesting examination included a 12-lead electrocardiogram (ECG), device interrogation, pacing and sensing threshold measurements, exclusion of myopotential interference, and evaluation of the intrinsic rhythm. The optimal pacing/sensing parameters determined for each patient were programmed and remained unchanged during testing. Pacing dependency was defined as a 2-s period of asystole or an escape rhythm at a rate ≤40 beats/min during pacing inhibition or during measurement of the capture threshold.
The exposure system consisted of a pair of rectangular, 120 × 140 cm, Helmholtz coils, 80 cm apart, mounted at the level of the patient's chest. A programmable source of alternative current was connected to the coils (model 6530, Chroma, Taipei-Hsien, Taiwan). Under the control of a computer using a LabView program (National Instruments, Austin, Texas), the source generated a 50-Hz magnetic field with a flux density programmable between 0 and 100 μT. The nominal voltage of the circuit was 16 V. Each coil consisted of 29 wires, 1.53 mm2in the cross-sectional area, receiving 5-V tension, generating 3-Amp current. The electrical field between the gates was on the order of 0.10 V/m. Three-dimensional calculation of the flux density with the EFC 400 software (Wandel and Goltermann, Eningen, Germany) confirmed the homogeneity of the magnetic field at the center of the induction volume (Fig. 1).The flux density, calculated as the total flux divided by the cross-sectional area of the volume through which it flows, was monitored through the exposure system by a sensor fixed on one of the gates at the level of the patient's chest. The room flux density was measured by a three-axial detector placed at a distance of 3 m away from the system.
No component of the system under tension was exposed, and the installation of the exposure system was approved by the local electrical safety commission. The patients were instructed to walk through the system at a normal pace, once parallel and once perpendicular to the gates, as well as stand at least 20 s inside the system. Thus, six exposures, three with and three without magnetic field generated, were randomly assigned to each patient, during each test. During the test, the time/density of the continuous signal of the magnetic field in the exposure system was monitored. The data collection included the frequency of the signal, voltage amplitude from the source, root mean square voltage, and current in the coils. The position of the patient, signal frequency, and flux densities in the room and in the gate were recorded every second in an Excel program application (Microsoft Corp., Seattle, Washington). A 12-lead ECG was continuously monitored using an independent computer-based ECG with an optical fiber connection to guarantee complete insulation of the patient from the computer. Attention was paid to select recordings free of motion or 50-Hz artifacts, which might have precluded a detailed analysis of the ECG. All tests were performed at a 100-μT maximum flux density. The test could be interrupted at any time, if necessary, or repeated to study its reproducibility. In case of interference, the control of the flux density between 0 and 100 μT was used to identify the lowest value causing the interference. Interrogation of the pulse generator was repeated after each test.
The data are presented as number and percentage of test, with the mean value ± SD and range.
A total of 250 tests were performed in 245 patients, five of whom had a second test after pulse generator replacement for battery depletion. The results are shown in Tables 1, 2, 3, and 4.⇓⇓⇓⇓Interference was observed in four (1.6%) of 250 tests. A mode switch from DDD to DOO pacing was recorded during the test in three patients with unipolar programming. Transient, asymptomatic, asynchronous dual-chamber pacing was recorded in two patients, one with unipolar atrial and ventricular sensing (Medtronic model 731) and the other with atrial unipolar sensing combined with bipolar ventricular sensing (Guidant model 1280). In a third patient, a switch to the asynchronous mode was followed by pacing inhibition (Fig. 2),resulting in complete atrioventricular (AV) block with profound bradycardia and lightheadedness (Medtronic model 7960). The lowest value inducing the mode switch was 45 μT. A mode switch was recorded in none of 153 tests of systems programmed in both atrial and ventricular bipolar sensing configuration, although during one test, transient ventricular pacing with a shorter than programmed AV delay was observed. This effect was caused by an interaction between the extracardiac signals and a specific algorithm used to confirm ventricular capture on a beat-by-beat basis. Bipolar atrial sensing at 0.75 mV was associated with bipolar ventricular sensing at 2 mV (St. Jude Medical, model 5376, AutoCapture). On post-test interrogation, reprogramming of no pulse generator was observed.
Interference by electrical appliances generating 50- or 60-Hz electrical or magnetic fields in close or direct contact with cardiac pacemakers is a known potential hazard (5). The main risk factors include device sensitivity, distance from the source of magnetic field, and field strength and orientation. In several clinical studies, the characteristics of the source of interference were poorly detailed or not monitored, and electrical and magnetic fields were often combined. Reprogramming of the sensitivity settings before testing, and variations in the exposure parameters lead to under- or overestimation of risks and consequences of interference. To our knowledge, a single clinical study has previously examined the risk of interference by magnetic fields. The fields were generated by 400-kV outdoor power plant substations located along roads. Interference was observed in one of 15 patients tested at the highest unipolar device sensitivity (3). The exposure system used in our study generated a continuously monitored, pure magnetic field, in the absence of any other electrical field. This system had been evaluated in preliminary tests with a 50-μT magnetic field, at 50- and 60-Hz frequencies (6). In our protocol, the 50-Hz frequency was the same as that of the European distribution of electricity, and the 100-μT flux density was at the recommended safety level for public exposure at 50 Hz (4). The simulation of the geometric effect in the field was included by orienting the device parallel or perpendicular to the gates. Implanted pacing systems form induction loops within which interference voltages may be induced by time-varying magnetic fields. In vitro studies have shown interference thresholds between 552 and 16 μT (root mean square) for magnetic fields at frequencies between 10 and 250 Hz (2).
Numerical simulations in millimeter-resolution, heterogeneous human body models have been performed to study the interference by 60-Hz magnetic fields with implanted unipolar pacemakers. Approximations derived from Faraday's law underscore the complexities of the induced current flowing through the human body, the length and placement of the leads with respect to the direction of the magnetic flow, and the inhomogeneous conductivity. Both the model and the input resistance of the pacemaker amplifier play critical roles in the results of these simulations. Estimated EMI thresholds under “worst case scenarios” were ∼40 μT for atrial electrodes at a sensitivity setting of 0.25 mV and 140 μT for ventricular electrodes at a setting of 0.75 mV (1). According to Faraday's law of induction, a left-sided unipolar permanent pacemaker is considered the most sensitive. In this configuration, the lead forms the largest inductive area, a semi-circular area ∼225 cm2, into which a magnetic field can induce a voltage. In bipolar systems, it was estimated that the field must be 17-fold larger to produce the same effect (7). The bipolar sensing configuration is the most protective against EMI. In recent pacemaker models, bipolar sensing is combined with self-adjustments enabling the settings of lower sensitivity levels than usual or nominal.
Our study shows a low incidence of interference by a high-density magnetic field in patients tested during routine follow-up visits, without changes in the programmed sensitivity settings or other pacing parameters made before the test. No interference was shown with bipolar programming, except for a clinically nonsignificant interaction with a specific capture threshold algorithm. In patients with unipolar sensing programming, the interference can cause sustained asynchronous mode reversion and pacing inhibition. Therefore, the risk of interference by a 50-Hz/100-μT magnetic field appears negligible in patients with bipolar sensing programming. AutoCapture function, which may be sensitive to EMI, should be disabled in patients who work in such environments. These clinical observations will help establish the specific limits of exposure to magnetic fields in patients with implanted pacemakers.
Continuous marker channel and intracardiac electrogram recordings allow a more accurate analysis of pacemaker behavior. These recordings were not used, because, in a preliminary study, direct interference by the magnetic field on the telemetry frequently interrupted the data transmission. Therefore, minor abnormalities on the surface ECG may have been missed.
Magnetic fields pulsed at power frequency caused an intermittent mode switch or pacing inhibition in a small percentage of patients with permanent pacemakers programmed in the unipolar sensing configuration. No device reprogramming was observed in this study. The overall incidence of interference was low with typical device programming.
We thank Jacques Lambrozo, MD, from the Department of Medical Studies, EDF-Gaz de France, and Robert Frank, MD, Cardiology Institute, Hôpital Pitié Salpétrière, Paris, France.
This work was supported by a grant from Le Réseau de Transport de l'Electricité and Electricité de France, Department of Medical Studies, Paris, France.
- Abbreviations and acronyms
- atrioventricular block
- electromagnetic interference
- Received August 22, 2004.
- Accepted October 18, 2004.
- American College of Cardiology Foundation
- Scholten A.,
- Silny J.
- Journal officiel des Communautés Européennes, L199/59, July 30, 1999
- Frank R.,
- Souques M.,
- Himbert C.,
- et al.