Medical Device Link.

Originally Published MEM Fall 2001

EMI RISK

Electromagnetic Interference in Medical Devices

An overview of EMI issues can help develop suitable electromagnetic compatibility requirements and other solutions to minimize risk to patients.

Kok-Swang Tan, Irwin Hinberg, and Jesuzette Wadhwani

Electromagnetic interference (EMI) has been responsible for a variety of medical-electronics malfunctions, which has raised concerns for patient safety. Between 1984 and 2000, Health Canada's Medical Devices Bureau received 36 reports of medical device malfunction attributed to EMI. Those malfunctions included four caused by wireless cellular phones. In two cases, EMI from electronic article surveillance (EAS) systems interfered with implantable cardiac pacemakers, and in one case, EAS was the possible cause of the premature failure of a pacemaker.^1,2 The bureau also investigated reports of interference from other radio-frequency sources, such as:

• Interference of an electrosurgical device with the electrocardiogram signals displayed on the monitor of an automated external defibrillator.
• Complete inhibition of the pacing signal of a pacemaker by a pulsating magnetic field from a video display terminal.
• Failure of the R-wave detection circuitry of a cardiac defibrillator in the presence of a simulated muscle artifact signal from an electrocardio- gram simulator.
• Interference of the line isolation system in an intensive care unit with the performance of a defibrillator.

Because of these reports, it was clear that guidelines were needed for the management of EMI within hospitals, especially in critical-care areas.^3,4

To identify emerging EMI issues and risk-management strategies, the bureau investigated the susceptibility of critical-care medical devices to various radio-frequency sources, including wireless telecommunications devices such as analog and digital cellular phones, two-way radios, personal communication service (PCS) systems, companion technology systems, family radio services (FRS) systems; wireless local-area network (LAN) systems; medical telemetry systems; and electronic security systems such as EAS systems, walk-through metal detectors, and handheld metal detectors. The risks were assessed for each type of system. Medical devices studied included therapeutic (29), monitoring (13), and diagnostic medical devices (10); implantable cardiac pacemakers (22); and implantable defibrillators (2).

Test Methodology

Portable and Permanent Medical Devices. Ad hoc tests were conducted to determine the susceptibility of portable medical devices to EMI from wireless telecommunications devices. Portable devices were tested in a vacant hospital room. Permanently mounted devices were tested on-site. Devices were tested under normal operating conditions. The test included four models of analog cell phones with a maximum power of 0.6 W, and four models of two-way radios with powers varying from 2 to 5 W. Susceptibility of the medical device to EMI from the wireless device was first tested at a distance of 3 m. If no EMI occurred at this initial setting, testing was repeated at shorter distances until either interference was noted or the telecommunications device was within 2 cm of the medical device.

At each distance, tests were performed with the antenna vertical and at different heights from the ground to the height of the medical device. All tests were repeated with the antenna horizontal and at the back, top, and both sides of the medical device. All distances were measured parallel to the ground from the base of the antenna on the telecom device to the face of the medical device. Next, the medical device was evaluated to determine whether it had been permanently damaged or reprogrammed.

Implantable Cardiac Devices. The in vitro testing setup for assessing the susceptibility of implantable cardiac pacemakers was similar to the procedure identified above. For these tests, a human torso simulator was used to simulate a human chest.5 The simulator consisted of a plastic box filled with 0.18% saline solution. The plastic box, which was placed horizontally on a table for testing wireless telecommunications devices, was placed vertically on a movable table for testing electronic security systems. Mounted on a grate, the pacemaker and its leads were submerged in saline.

Figure 1. The experimental setup for studying EMI effects from radio-frequency sources on implantable cardiac pacemakers.

The schematic diagram in Figure 1 shows the experimental setup. Signals obtained from two stainless-steel plates were amplified with a differential amplifier, and the signal from the amplifier was displayed on a storage oscilloscope. Simulated electrocardiographic (ECG) signals from a patient simulator were displayed on an ECG monitor. The amplified ECG signals from the monitor were applied to two stainless-steel plates mounted perpendicular to the first two. These plates applied simulated ECG signals to inhibit the pacemaker. Implantable defibrillators were tested using the same setup.

Any changes in the pacing signals, such as changes in the amplitude and the pulse duration of the signals, the intervals between the atrial and ventricular pulses, and the pacing frequency, were recorded. These changes determined whether the signal from the radio-frequency sources affected the pacing outputs to induce a total, partial, or prolonged inhibition.

After establishing the threshold amplitude of the amplified ECG signal, the output amplitude of the amplified ECG signal was set at double its threshold amplitude. The effect of doubling the amplitude could be to change the pacemaker's operating mode to an asynchronous mode with either a rate different from the simulated ECG signals or to a rate synchronous with the same pacing frequency as the simulated ECG signal. If reactivation occurred, the pattern of pacing (either regular or irregular) was recorded, which determined whether EMI could reactivate the pacemaker's pacing signal after it had been inhibited by the simulated ECG signal.

Wireless Telecommunications Devices. Each wireless telecommunications device was tested at different points on the grate with the antenna of the wireless device oriented either parallel or perpendicular to the long axis of the simulator. For analog and digital cellular phones and FRS systems, actual communications were transmitted either by keying out or calling in to the wireless device. Only keyed-out communications were programmed for the PCS system.

Security Systems. Both static and dynamic tests were conducted for walk-through security systems. The simulator was placed on a turntable, which was mounted on a wooden table with rigid rubber wheels. The pacemaker's header was 122 cm above ground. Static tests simulated a person standing with a shoulder against the transmission panel of a security system (perpendicular orientation). Rotating the turntable 90° simulated a person's back against the panel (parallel orientation). Tests for inhibition and reactivation were conducted at various locations along the transmission panel and various turntable angles. Dynamic tests were carried out by rolling the table at a speed of 20 cm/sec to simulate a person walking through a security system. Tests were repeated for the parallel orientation.

A handheld metal detector was swept slowly in front of the simulator five times to simulate a pacemaker patient being scanned by security personnel. The tests included three models of EAS systems from two manufacturers, six walk-through metal detectors from three manufacturers, and 13 handheld metal detectors from nine manufacturers.

Results

More than 60% of the medical devices tested were affected by interference from a wireless telecommunications device. Of these, 64% (32 out of 50) were affected by two-way radios, and 26% (13 out of 50) were affected by analog cell phones. Two-way radios caused the greatest interference at the farthest distances. Cell phones affected devices much less and often occurred only at very close distances. Two-way radios caused some devices to malfunction at 3 m, whereas analog cellular phones did not cause malfunction at distances greater than 1 m. Each effect was assigned one of the following fault codes:

• A = Visible and/or audible alarm with device stoppage.
• B = Visible and/or audible alarm with no device stoppage.
• C = Readout errors with no change in operation.
• D = Measured readout variation causing change in device operation.
• E = Audio indicator distortion.
• F = Device reboots or powers down by itself.
• G = Loss of input being measured.
• H = Distortion present in device measurement display (waveforms).
• I = Device malfunction, no alarms.
• J = Display malfunction.
• K = Device needed to be manually reset to continue proper operation.
• L = Device changes operating mode.
• M = Recorder malfunction.
• N = Alarm malfunction.
• O = No observed effect.
• P = Device could not be approached from that direction.

The most common faults were C, D, and F. The errors caused by cellular phones were not as severe as those that occurred with the use of two-way radios.

Table I indicates that the wireless telecommunications devices interfered with the operation of the majority of medical devices tested. However, some devices were affected only at very close ranges and at specific frequencies. The severity of EMI depended on the following:

• The distance from the wireless telecommunications device to the medical device.
• The distance from the wireless telecommunications device to the transmission tower base station that determined the power output of the wireless device.
• The frequency of the wireless device.
• The transmitting time in some cases.
• The shielding of the medical device and cables.

Figure 2 illustrates susceptibility curves for the six medical devices that were the most susceptible to EMI from wireless devices. Four of the devices failed to meet the International Electrotechnical Commission electromagnetic compatibility requirement of 3 V/m.⁶

Figure 2. Susceptibility tests of six medical devices.

The results also indicated that cell phones cause greater interference when establishing a connection. Once established, the severity often diminishes. The difference could be attributed to the fact that cellular phones adjust power output depending on the distance from the operating site to the base station.

Wireless LAN and Medical Telemetry Systems

A wireless LAN system operating at 2.42 GHz generated an electric field strength of 0.1 V/m at 1 m from the antenna of the system. The background electric field strengths at each test site were below 0.1 V/m, except near an elevator outside the operating rooms, where the electric field strength varied from 0.15 to 0.25 V/m.

The bureau tested 106 medical devices located in three hospitals, one with 300 beds and two with 800–900 beds. Devices were assessed for susceptibility to EMI from two wireless LAN systems and one medical telemetry system. When the wireless LAN system was in the standby mode and within 10 cm of three models of handheld Doppler units, the Doppler units produced periodic high-pitched beating sounds, which could be misinterpreted as sounds from a patient.⁷ The periodic sounds changed to random static noise when the wireless LAN system was transmitting data. In addition, transmission quality deteriorated in the colonoscopy room, possibly because the room was shielded with lead. A hardwired connection from the wireless LAN prevented deterioration of transmission quality. A wireless LAN system employing very low field intensities could be used safely near medical devices. None of the devices was affected by a medical telemetry system operating at 466 MHz.

Wireless Telecommunications Systems

Nine of the 20 pacemakers tested were susceptible to EMI from nearby digital cellular phones.⁵ However, interference was generally not present when the cell phone was more than 15 cm from the pacemaker (see Figure 3). Dual-chamber pacemakers were more susceptible to interference than single-chamber models. All four types of digital cell phones decreased the output pulse rate of two single-chamber pacemakers and two dual-chamber pacemakers. The cell phones also induced rhythmic pacing in three single-chamber pacemakers and four dual-chamber pacemakers when the output pacing was inhibited by an external ECG signal. EMI caused a tenfold increase in the monitoring peak amplitudes of the pacing output from a single-chamber and a dual-chamber pacemaker from the same manufacturer. The rate of interference (3.4%) is consistent with results reported in other studies (see Table II). It is important to note that EMI generated by wireless devices did not reprogram the pacemakers, and interference ceased when phones were turned off. Analog cell phones, digital PCS systems, and FRS systems did not interfere with pacemakers. The results of this study indicate that pacemaker patients should use analog or PCS phones. Hayes et al. also found that analog cellular phones are safe for patients with pacemakers.⁸

Figure 3. Three-dimensional spatial effects from digital cellular phones on pacemakers. The 3-D contour shows the interference locations.

Dual-chamber pacemaker models, with and without filters, were tested to investigate the effectiveness of filtering circuitry. No EMI was observed in the model with filters; however, transient inhibition was observed in the model without filters. A clinical study by Carillo also noted no interference to pacemakers with filters.⁹ Pacemaker manufacturers should use filtering technology to reduce the susceptibility of pacemakers to EMI.

To determine the vertical distances needed to induce EMI between a wireless system and a pacemaker, additional testing was done on the pacemakers susceptible to EMI. A digital cellular phone was placed on wooden blocks above the simulator grate, and the distance was measured when an inhibition or a reactivation was observed. Figure 3 shows a three-dimensional graph of EMI on nine pacemakers. The average maximum vertical distance at which interference was observed was 3.4 cm above the pacemaker. However, in one case, interference occurred at 40 cm from the pacemaker. In all cases, EMI occurred only when the phone was transmitting.

Electronic Article Surveillance Systems

The bureau also studied the susceptibility of 22 pacemakers and two implantable defibrillators to EMI from three types of EAS systems (pulsed magnetic field, continuous magnetic field, and sweep-frequency magnetic field.^10,11 During static tests, pulsed-magnetic systems decreased the electrical output pulses of eight of the pacemaker models. Sweep-magnetic systems affected five pacemaker models, but continuous-magnetic systems did not affect any pacemakers (see Table III). The decrease in the pacing frequencies of some pacemakers could be a concern for patients fully dependent on pacemakers.

Such patients could experience dizziness or collapse if pacing ceases for more than 3 seconds. When the pacemaker output was inhibited by an external ECG signal, pulsed-magnetic systems induced rhythmic pacing on 15 pacemakers at distances up to 33 cm from the transmission panel. The pulsed-magnetic systems affected 12 pacemakers at up to 18 cm. This asynchronous pacing could compete with the normal sinus rhythm and therefore induce arrhythmia. In addition, a bipolar lead configuration appeared to be less susceptible to interference from pulsed-magnetic systems than a unipolar lead configuration. The same EMI effects occurred both inside and outside the transmission panels of the EAS gate. No interference occurred when the simulator was rolled through any EAS system during the dynamic tests. The two defibrillators were also unaffected by any EAS systems. However, McIvor and Mathew et al. have reported that patients with implantable defibrillators experienced inappropriate firing when near EAS systems.^12,13 No EAS system caused any permanent damage or reprogramming of the pacemakers or defibrillators. Pacemaker patients, therefore, should not stop within 33 cm of either side of the transmission panel of an EAS.

Walk-Through and Handheld Metal Detectors

None of the 13 handheld metal detectors interfered with the pacemakers and the implantable defibrillators tested.¹² One walk-through metal detector decreased the pacing frequencies of five single-chamber and three dual-chamber pacemakers during the inhibition tests. Ten single-chamber and four dual-chamber pacemakers in-duced reactivation. The other five walkthrough metal detectors caused interference with only two single-chamber pacemakers during inhibition and reactivation tests. No interference was observed when the simulator was rolled through the walk-through metal detectors during the dynamic tests. It appears that walk-though metal detectors may affect some pacemaker models if patients remain near the detector for longer than 2 seconds. Only one walk-through metal detector stopped defibrillation on one defibrillator when a ventricular fibrillation was simulated. None caused reprogramming or permanent damage to any medical devices. Both pacemaker and defibrillator patients should avoid stopping within 33 cm of either side of the metal detector's transmission panel.

Conclusion

Although many medical device manufacturers have addressed some of these EMC issues, much progress needs to be made. In addition to educating device users, manufacturers and regulators must continue to develop EMC standards to minimize such interference.

The exposure to electromagnetic environment, as well as the frequency, location, orientation, and design of a device, all influence whether and how a device might be affected by EMI. In practice, it is impossible to stop electromagnetic energy completely at its source, and society has become dependent on the convenience of instant communication. Some medical devices themselves emit electromagnetic energy, and simultaneous use of such devices can also cause EMI problems.

Studies of EMI susceptibility of medical devices consistently report that wireless telecom devices can cause a high percentage of medical devices to malfunction. However, a relatively small number of EMI incidents have been reported to regulatory agencies in both Canada and the United States. Some underreporting can be attributed to a lack of awareness of EMI as the cause of the malfunction. It has been suggested that hospitals should develop policies for addressing the use of wireless devices in their facilities. Participants of the Round Table on EMC convened by Health Canada in September 1994 unanimously agreed that cell phones should not be totally banned in hospitals, but that policies governing their use should be established.

The results of this study confirm that the very-low field intensities generated by wireless LAN systems do not interfere with medical devices in hospitals. These wireless systems are acceptable for use in hospitals, especially in view of the benefits of obtaining real-time access to patients' medical information. Nevertheless, each new wireless system should be tested on potentially susceptible devices by the hospital before being put into general use. One of the conclusions of the Canadian Task Force on EMC in Health Care was that "the potential for EMI is minimized by using RF sources having the lowest possible transmission power."¹⁴ The task force recommended replacing high-power sources with lower-power ones.

Furthermore, these recent studies confirm that EMI from nearby digital cellular phones could affect the operation of some implanted cardiac pacemakers. The greatest risk of interference occurs when the antenna of the phone is close to the pacemaker. The chances that EMI from digital cellular phones would produce a life-threatening situation are minimal. However, the data support the recommendations in the Health Protection Branch's Medical Device Alert No. 108.¹⁵

Interference by security systems such as EAS systems and walk-through metal detectors on pace- makers may pose a potential risk, depending on the model of pacemaker, the programmed mode, and the type of the security system.

Security system manufacturers should consider labeling transmission panels to identify the presence of a security system and to advise pacemaker and defibrillator patients to walk through such systems at a normal pace. EAS systems should not be installed in the vicinity of cash registers in such a way that pacemaker-dependent patients must stand within 33 cm of the EAS system while waiting.

Many EMI malfunctions could be prevented through education. Health Canada places a high priority on providing information to the public. The department sends Medical Device Alerts and It's Your Health bulletins to healthcare professionals to warn of serious problems and to provide recommendations to manage the associated risks.

It is imperative that manufacturers design medical devices to be immune to electromagnetic fields up to 10 V/m for life-support medical electrical equipment and 3 V/m for non-life-support medical electrical equipment, as proposed in international standards. Meeting these standards would greatly reduce the potential hazards of EMI. Specific EMC standards for implantable cardiac pacemakers and defibrillators are currently being drafted by ISO technical committee 150, subcommittee 6, working groups 1 and 2.¹⁶

Kok-Swang Tan is a research scientist, Irwin Hinberg is the head of the Criteria and Assessment Section, and Jesuzette Wadhwani is a medical technologist and research assistant. They are with the Medical Devices Bureau, Therapeutic Products Directorate, Health Canada (Ottawa, ON, Canada).

References

1. Medical Devices Incident Report Database, Medical Devices System, Medical Devices Bureau, Therapeutic Products Programme, Health Canada.

2.KS Tan, et al, (edited), in Proceedings of the Round Table Discussion on Electromagnetic Compatibility in Health Care (Ottawa, ON, Canada, August 1995).

3.KS Tan and I Hinberg, "Investigation of Electromagnetic Interference with Medical Devices in Canadian Hospitals," in Proceedings of a Workshop on Electromagnetics, Health Care and Health (Montreal, September 19–20, 1995).

4."Electromagnetic Interference in Medical Devices: Health Canada's Past and Current Perspectives and Activities," in Proceedings of the IEEE International Symposium on Electromagnetic Compatibility (Montreal, IEEE EMC Society, 2001).

5.KS Tan and I Hinberg, "Can Wireless Communication Systems Affect Implantable Cardiac Pacemakers? An In Vitro Laboratory Study," Biomedical Instrumentation Technology 32, no. 1 (1998): 18–24.

6.IEC 60601-1-2 draft 2nd ed., "Medical Electrical Equipment Part 1: General Requirements for Safety. –2. Collateral Standard: Electromagnetic Compatibility—Requirements and Tests," International Electrotechnical Commission, Geneva.

7.KS Tan and I Hinberg, "Effects of a Wireless LAN System, a Telemetry System and Electrosurgical Devices on Medical Devices in a Hospital Environment," Biomedical Instrumentation Technology 34, no. 2 (2000): 115–118.

8.DL Hayes et al., "Interference with Cardiac Pacemakers by Cellular Phones," New England Journal of Medicine 336, no. 21 (1997): 1473–1479.

9.RG Carillo et al., "Electromagnetic Filters Impede Adverse Interference of Pacemakers by Digital Cellular Phones," Journal of the American College of Cardiology, (1996): Abstract 901-22, 15A.

10. KS Tan and I Hinberg, "Can Electronic Article Surveillance Systems Affect Implantable Cardiac Pacemakers and Defibrillators?" Pacing and Clinical Electrophysiology 21, no. 4 (1998): 960.

11.KS Tan and I Hinberg, "A Laboratory Study of Electromagnetic Interference Effects from Security Systems on Implantable Cardiac Pacemakers," in Proceedings of the XXVI General Assembly International Union of Radio Science on Electromagnetic Interference with Medical Devices (Toronto, General Assembly Union of Radio Science, 1999).

12.ME McIvor, "Environmental Electromagnetic Interference from Electronic Article Surveillance Devices: Interactions with an ICD," Pacing Clinical Electrophysiology 18, no. 12 (1995): 2229–2230.

13.P Mathew et al., "Interaction Between Electronic Article Surveillance Systems and Implantable Defibrillators: Insights from a Fourth-Generation ICD," Pacing Clinical Electrophysiology 20, no. 11 (1997): 2857–2859.

14.S Segal, et al., "Recommendations for Electromagnetic Compatibility in Health Care," in Proceedings of the Canadian Medical and Biological Engineering Conference 22, (1996): 22–23.

15.Medical Device Alert No.108, Digital Cellular Phone Interference with Cardiac Pacemakers, Health Canada, Ottawa, ON, Canada, November 6, 1995.

16.KS Tan and I Hinberg, "Susceptibility of Medical Devices to Radio-Frequency Electromagnetic Fields," in Proceedings of the Advance Clinical Engineering Workshop, Santo Domingo, Dominican Republic, March 13–17, 2000.

Copyright © 2001 Medical Electronics Manufacturing