Electromagnetic Interference With Implantable Cardioverter-Defibrillators at Power Frequency
An In Vivo Study
Background—The number of implantable cardioverter-defibrillators (ICDs) for the prevention of sudden cardiac death is continuing to increase. Given the technological complexity of ICDs, it is of critical importance to identify and control possible harmful electromagnetic interferences between various sources of electromagnetic fields and ICDs in daily life and occupational environments.
Methods and Results—Interference thresholds of 110 ICD patients (1-, 2-, and 3-chamber ICDs) were evaluated in a specifically developed test site. Patients were exposed to single and combined electric and magnetic 50-Hz fields with strengths of up to 30 kV·m−1 and 2.55 mT. Tests were conducted considering worst-case conditions, including maximum sensitivity of the device or full inspiration. With devices being programmed to nominal sensitivity, ICDs remained unaffected in 91 patients (83%). Five of 110 devices (5%) showed transient loss of accurate right ventricular sensing, whereas 14 of 31 (45%) of the 2- and 3-chamber devices displayed impaired right atrial sensing. No interference was detected in 71 patients (65%) within the tested limits with programming to maximum sensitivity, whereas 20 of 110 subjects (18%) exhibited right ventricular disturbances and 19 of 31 (61%) subjects exhibited right atrial disturbances.
Conclusions—Extremely low-frequency daily-life electromagnetic fields do not disturb sensing capabilities of ICDs. However, strong 50-Hz electromagnetic fields, present in certain occupational environments, may cause inappropriate sensing, potentially leading to false detection of atrial/ventricular arrhythmic events. When the right atrial/right ventricular interferences are compared, the atrial lead is more susceptible to electromagnetic fields.
Exogenous electric and magnetic fields (EMFs) from sources such as high-voltage power lines, substations, electronic article surveillance systems, or electrical appliances induce noise signals in the human body. They superimpose intrinsic heart signals and may lead to electromagnetic interference (EMI) with active implantable medical devices such as cardiac pacemakers or implantable cardioverter-defibrillators (ICDs).
Clinical Perspective on p 450
First reports of EMI with cardiac implants were published in the early 1960s.1 During the last decades, many different studies dealing with EMI have been published; however, to date there is no conclusive evidence as to which sources of extremely low-frequency EMFs may disturb cardiac pacemakers or ICDs.
New indications for ICD therapy, especially the implantation for primary prevention of sudden cardiac death, have substantially increased the number of patients carrying an ICD, thus emphasizing the need for further investigation and regulation. In Germany alone, there were 21 609 ICD implantations in 2006, increasing to 42 261 in 2012.2,3 Many of these devices were implanted in relatively young patients possibly still working with the risk of strong field exposure in specific occupational environments (eg, technician in a power plant). In 2011, 51.1% of all patients with first ICD implantation were <70 years of age and 25.5% were <60 years of age in Germany.4 Additionally, the number and types of EMF sources have likewise risen in daily life and occupational environments over the past 2 decades. EMI with ICDs may cause inadequate oversensing with subsequent inappropriate shock delivery, inhibition of pacing, or switch to an asynchronous noise mode. Of note, inappropriate shock delivery seems to carry an increased risk for overall survival5 and cause psychological distress.
Safety guidelines of the European Union,6–8 the American National Standards Institute (ANSI),9 and the International Commission on Non-Ionizing Radiation Protection10 for the protection of humans exposed to EMFs do not include patients with medical devices such as ICDs (Table 1). Nevertheless, several product standards for manufacturers provide test methodologies (so-called benchmark tests) to evaluate the electromagnetic compatibility performance of ICDs.11–14 However, these standards achieve electromagnetic compatibility only to a certain degree. It is important to understand that EMI may occur despite conformance of cardiac implants to the specific product standards and the conformance of sources of EMFs to the human exposure safety guidelines according to the ANSI/Association for the Advancement of Medical Instrumentation standard PC69.14 There is a lack of comprehensive data, especially on in vivo exposure, to close the current gap in knowledge about the extent to which patients with cardiac implants may be influenced by extremely low-frequency EMFs.
The objective of the present study was to provide sound data on the exposure of ICD patients to extremely low-frequency EMFs to overcome the existing uncertainty among patients and physicians. In a clinical in vivo provocation study, ICD patients were exposed to single and combined 50-Hz EMFs of up to 30 kV·m−1 (electric field strength) and 2.55 mT (magnetic flux density). These are the maximum occupational limits in Germany covering EMFs in the vicinity of high-voltage power lines, power installations, or other power-operated machines.15 We systematically determined the interference thresholds of patients’ ICDs, that is, the first occurrence of a sensing failure under worst-case conditions (eg, maximum sensitivity).
The study design was approved by the institutional review committee for Human Research of University Hospital RWTH Aachen (www.clinicaltrials.gov; identifier NCT01626261). It consisted of 2 parts: (1) benchmark tests, that is, computer-based tests with ICDs to develop and validate the method of the provocation study, and (2) a provocation study, that is, a clinical in vivo study with ICD patients to determine the individual interference thresholds.
The purpose of the benchmark tests was to determine whether the interference threshold is independent of the duration of exposure to validate the method of short-term exposure in the in vivo provocation study. A computer-generated intracardiac electrogram, corresponding to the European product standard EN 45502-2-2,11 superimposed with a 50-Hz sinusoidal noise signal was fed into the pace/sense channel of the ICD to be tested. The injected 50-Hz sinusoidal noise signal simulated the EMF exposure used in the provocation study. Interference thresholds of different ICDs were determined and compared under short-term exposure (1.5 seconds, ≈2 consecutive heartbeats) and long-term exposure (30 seconds). The ICD parameters were set to nominal settings but maximum sensitivity. The reaction of the ICD was monitored by standard programming devices. The amplitude of the 50-Hz sinusoidal noise signal was increased successively until the first sensing failure of the device occurred (ie, the interference threshold) or the maximum amplitude (20 mV peak to peak) was reached. A total of 15 different ICD models were tested (Table 2). ICDs were previously explanted in patients as a result of battery depletion, device infection, or upgrade to a different system.
In the provocation study, ICD patients were systematically exposed to EMFs of different intensities to define thresholds of EMI at 50 Hz. For safety reasons, sequences of field exposure in patients were limited to a maximum of 2 consecutive heartbeats (short-term exposure), and ICD therapies for ventricular tachycardia (VT) and ventricular fibrillation (VF) were switched off during the investigation.
During the period of September 2009 to December 2012, all patients presenting to the outpatient pacemaker/ICD clinic of our department were screened for the study. Of 1983 patients consecutively requested for routine ambulatory ICD follow-up, 386 patients met the inclusion/exclusion criteria, and 110 gave written informed consent. Inclusion criteria were age between 18 and 75 years and device implantation >4 weeks previously. Exclusion criteria were pacemaker dependency, hyperthyroidism, ineffective oral anticoagulant therapy in case of atrial fibrillation, serum electrolyte disorders, clinically manifest infection, myocardial infarction <30 days, and pregnancy.
Pretest examination included a 12-lead ECG, device interrogation, and analysis of blood samples (electrolyte levels and coagulation). Body measurements (height, weight, thorax circumference, shoulder width) and information about the implanted system (manufacturer and model of the device and leads, chest X-ray) were documented.
Follow-up examination, immediately after the test and again after 4 weeks, included a 12-lead ECG and device interrogation. No device defects or software resets were seen. Pacing thresholds remained unchanged at the follow-up visits.
All 110 patients who consented to participate in the study were included to obtain a comprehensive picture of interference thresholds of ICDs. Single-chamber ICDs were implanted in 79 patients, dual-chamber ICDs in 16 patients, and 3-lead ICD systems (cardiac resynchronization therapy–defibrillator [CRT-D]) in 15 subjects. One patient (P090) with a dual-chamber ICD was programmed to the VVI mode because of an atrial lead defect. This patient was included as single-chamber ICD. Table 3 shows the characteristics of the 110 patients.
A computer-controlled test system was developed to continuously monitor EMF generation and to record the patient’s surface ECG signals in real time. Standard programming devices were used to register intracardiac electrograms and marker channels during the entire examination. The system aligned the phasing between the magnetic and electric fields so that in combined field settings a maximum disturbance at the input of the ICD was always ensured. Figure 1 shows a scheme of the technical test setup.
Electric Field Generation
Generation of strong homogeneous electric 50-Hz fields requires special constructions to comply with safety regulations and to ensure homogeneous fields.16 The geometric constraints of the laboratory did not allow this. However, the induced current distribution in the thorax by exogenous electric fields can be reproduced by direct current injection. See the Methods section and Figure I in the online-only Data Supplement for validation. The body current that would be induced by a vertically oriented exogenous electric field is given according to the Deno formula.17 The system injected this defined body current between the neck and feet of the patient. This procedure has been described in detail elsewhere.18 The test setup can produce body currents up to an equivalent exogenous field strength of 30 kV·m−1 (ie, the occupational limit in Germany).15
Magnetic Field Generation
Homogeneous magnetic 50-Hz fields were generated by a vertical Helmholtz coil setup with a diameter of 180 cm and a distance of 90 cm. The Helmholtz pair of coils was loaded by a power amplifier (Vortex 6, Camco) and a signal generator (33220A, Agilent Technologies). In this way, homogeneous 50-Hz magnetic fields of up to 2.55 mT flux density (ie, the occupational limit in Germany)15 can be produced.
Field exposure was applied for 2 consecutive heartbeats. Each sequence of exposure was triggered to the R wave of the surface ECG to start exposure during the blanking time of the device, thus preventing interferences caused by field initiation.
The objective was to determine the lowest interference thresholds for ICD patients, that is, the first occurrence of a sensing failure, in either single or combined EMFs. To identify reliable lowest thresholds, worst-case conditions were required and defined as follows:
Thorax perpendicular to the orientation of the homogeneous magnetic field. According to the Faradays law, the induced noise signal voltage is proportional to the induction area. The induction area becomes maximum when magnetic fields act perpendicularly to the frontal plane of the thorax.19
Full inspiration. Exposure of electric fields by full inspiration results in an increase in the noise signal voltage in the body and hence a potential decline in the interference threshold. It is supposed that the large volume of insulating air causes a higher local current density in the area between lung and chest skin.18
Sustained pacing, that is, a pacing rate set to a higher value than the intrinsic heart rate and, for a 2- or 3-chamber device, a foreshortened AV delay. The adjustment of the sensitivity can only partly be programmed. ICDs automatically adjust their atrial and ventricular sensitivity thresholds after sensed and paced events, and only a few parameters, for example, the sensitivity value, can be set. Depending on the intracardiac electrogram P/R potential, device settings, and manufacturer, the sensitivity may be higher after pacing than after sensing, as explained in device manuals.
Maximum sensitivity. ICDs were programmed to the highest obtainable sensitivity because the interference threshold is coupled with sensitivity settings.20 Sensitivity has the greatest impact on interference thresholds of cardiac implants.21
The examination was conducted at maximum and nominal sensitivity, applying worst-case conditions stepwise.
First Run: Maximum Sensitivity
ICDs were set to maximum sensitivity, and interference thresholds were determined for single and combined EMFs. Then, a second worst-case parameter was added: The pacing rate was adjusted to ensure continuous atrial/ventricular pacing, and the previously determined thresholds were reassessed. Finally, the influence of respiration on the thresholds was analyzed by repeating the exposures while the patient was at full inspiration.
Second Run: Nominal Sensitivity
ICDs were set to nominal sensitivity but maintaining the other worst-case conditions. Nominal sensitivity means that the preset sensitivity is programmed by the treating physician. Thus, nominal sensitivity and maximum sensitivity can be equal (eg, in patients with preexisting low R potential). Interference thresholds were determined for single and combined EMFs at nominal sensitivity. After programming of AV sequential/right ventricular pacing and investigation of the influence of full inspiration, thresholds were again determined.
At each run and condition, field strengths were increased stepwise until the individual thresholds were found or maximum field values (30 kV·m−1/2.55 mT) were reached. For validation, the exposure of the determined threshold was repeated twice. The strategy of increasing the field strength was based on a binary decision tree, permitting precise determination within a maximum of 6 steps.
Statistical analysis was performed with MATLAB (MathWorks). Unless otherwise specified, data are expressed as mean and standard deviation.
The interference thresholds of 15 different ICD models were determined, which were also part of the provocation study. The first sensing failure occurred at noise signal amplitudes between 0.14 and 1.2 mV. Dual-chamber or CRT-D systems showed lower interference thresholds than single-chamber ICDs because of the higher sensitivity of the atrium channel. The interference thresholds of the 1.5-second exposure were almost identical to the thresholds of the 30-second exposure (Table 2). The slight differences can be explained by the standard measurement uncertainty. The reaction/dysfunction of ICDs at the first sensing failure (ie, the interference threshold) remained the same, independently of the duration of exposure (either short-term or long-term exposure). Thus, these results indicate that the interference thresholds obtained in the in vivo provocation study are also valid for permanent field exposure, assuming that the other conditions remain constant. In conclusion, the in vivo provocation study allows a general risk assessment of susceptibility to EMI even if the patients were only exposed short term (2 consecutive heartbeats).
Of note, the interference thresholds determined in the benchmark tests cannot be linked directly to the interference thresholds obtained from the provocation study because of the missing patient- and lead-related effects.
At maximum sensitivity, no interference during EMF exposure occurred in 71 of 110 implanted devices (64.5%). The noise signal provoked inadequate ICD responses in 19 patients in the atrial channel and in 20 individuals in the ventricular channel (Figure 4).
Programmed at nominal sensitivity, no disturbance occurred in 91 of 110 devices (82.7%). In 14 of these 19 ICDs with EMF interference, the atrial channel was affected, whereas in 5 patients, interference occurred in the ventricular channel (Figure 4).
At interference thresholds, oversensing in both the atrial and ventricular channels was the type of the first sensing failure. None of the tested devices primarily switched into noise mode when the first oversensing occurred. One CRT-D showed inhibition of left ventricular pacing (P077).
The percentage of disturbed ICDs per year of market release showed no dependency on susceptibility to 50-Hz EMFs (data not shown). Hence, newer ICD models seem not to be less susceptible, although the number of implants per year of release was not sufficient for a statistical validation.
Right Atrial Disturbances
Atrial interferences were detected in 19 of 31 (10 dual-chambers; 9 CRT-Ds) 2- or 3-chamber ICDs (61.3%) at maximum sensitivity in combined fields of up to 30 kV·m−1 (electric field strength) and 2.55 mT (magnetic flux density). Figure 2 shows an example of EMI.
In single electric or magnetic field applications, 13 devices showed interference: 6 (19.4%) in the electric field (3 dual chambers; 3 CRT-Ds) and 7 (22.6%) in the magnetic field (5 dual-chambers; 2 CRT-Ds).
At nominal sensitivity, 14 of 31 ICDs (45.2%; 9 dual-chambers; 5 CRT-Ds) showed an atrial disturbance in combined fields of up to 30 kV·m−1 and 2.55 mT.
In single-field applications, only 3 ICDs (9.7%; 2 dual-chambers; 1 CRT-D) could be disturbed in magnetic fields and 1 ICD (3.2%) in electric fields.
The lowest atrial interference thresholds were 15 kV·m−1 in single electric fields (atrial sensitivity, 0.18 mV), 0.45 mT in single magnetic fields (atrial sensitivity, 0.2 mV), and 3.5 kV·m−1/0.3 mT in combined EMFs (atrial sensitivity, 0.4 mV). Details are given in Table I in the online-only Data Supplement.
Right Ventricular Disturbances
Ventricular interferences were detected in 17 of 110 ICDs (15.5%) at maximum sensitivity in combined fields of up to 30 kV·m−1 and 2.55 mT. The disturbed ICDs were 16 single-chamber ICDs and 1 CRT-D. An example is depicted in Figure 3.
During single-field application, only 10 ICDs (9.1%) were disturbed in magnetic fields (8 single-chamber; 2 CRT-Ds) and none in electric fields. One CRT-D (P077) showed an interference in the left ventricular channel. It was the only left ventricular disturbance elicited in the study; however, it should be noted that only Biotronik and Guidant/Boston Scientific CRT-Ds provide left ventricular intracardiac electrograms.
At nominal sensitivity, ventricular oversensing of combined fields was observed in 5 of 110 ICDs (4.5%; 4 single-chamber; 1 dual-chamber). Single magnetic field exposure interfered with 1 ICD (0.9%). Single electric field exposure did not interfere with ICDs at nominal sensitivity.
The lowest ventricular interference thresholds were >2.55 mT in single electric fields, 0.6 mT in single magnetic fields (ventricular sensitivity, 0.18 mV), and 16.4 kV·m−1/1.8 mT in combined EMFs (ventricular sensitivity, 0.18 mV). Details are given in Table II in the online-only Data Supplement.
Interference Thresholds in Relation to Limit Values
The interference thresholds of the disturbed ICDs at maximum and nominal sensitivity within the range of the limit values set by the European Union and the ANSI/Institute of Electrical and Electronics Engineers (IEEE; Table 1) are shown in Figure 5.
At maximum sensitivity, 39 ICDs could be disturbed in single and combined fields up to the tested limits (30 kV·m−1/2.55 mT). Of these, no interference occurred within the limit values set by the European Union for the general public, and only 1 device could be disturbed within the set occupational limits.
With respect to the limits defined by the ANSI/IEEE C95.6 guideline, the interference thresholds of 3 ICDs with atrial interference and 1 ICD with ventricular interference were in the range for the general public. Within the occupational limits of this guideline, the ICDs of 21 patients could be disturbed: 10 atrial and 11 ventricular disturbances respectively.
Focusing on nominal sensitivity, 19 ICDs could be disturbed within tested limits; of these, 3 ICDs with atrial interference were within the limits of the ANSI/IEEE C95.6 guideline for general public, and 6 ICDs were within the limits for occupational exposure (4 atrial and 2 ventricular disturbances). At nominal sensitivity, no interference could be elicited within the European Union limit values in either the general public or the occupational limits.
The main findings of the present study are (see also Table 4) as follows:
Interference of EMFs with ICDs occurred in 17.3% at programmed nominal sensitivity and in 35.5% at maximum sensitivity within the tested limits.
Interference of EMFs with the ventricular channel occurred in 4.5% of ICDs at nominal sensitivity and in 15.5% at maximum sensitivity.
EMF interference with the atrial channel occurred in 45.2% at nominal sensitivity and in 61.3% at maximum sensitivity.
No interference occurred within the European Union limit values set for the general public, and only 1 device could be disturbed in the atrial channel within the occupational limits.
Within the limit values of the United States (ANSI/IEEE), EMI with the atrial and the ventricular channel occurred in 4% (general public limits) and 19% (occupational limits) of all patients.
Active pacing of the ventricle increased the susceptibility to EMI by 91% of all tested ICDs.
The EMF-Portal (www.emf-portal.org), the most comprehensive scientific literature database on the effects of EMFs, currently reveals ≈300 publications on EMI with cardiac implants. Although 189 studies have investigated EMI in the low-frequency range (including direct current), only 47 publications have dealt with the power frequency range (50/60 Hz). However, many of these 47 publications were conducted on various numerical and physical models (eg, see References 22–26). Additionally, there have been a number of case studies (eg, References 27 and 28) or retrospective observational studies (eg, References 29 and 30). Another group of publications comprises investigations on EMI caused by medical electrical equipment working in the 50/60-Hz range.31–33 Even though the first evidence of EMI appeared in the early 1960s,1 to date, there have been only 4 clinical studies21,34–36 with patients bearing a cardiac implant under standardized or controlled exposure conditions in the 50/60-Hz power frequency range. Nevertheless, provocation studies were recommended in numerous previous studies.24,33,35,37 Trigano and coworkers34 showed in a large in vivo study of cardiac pacemaker patients that single magnetic fields pulsed at power frequency are able to cause an inappropriate mode switch and pacing inhibition in unipolar lead configuration. Bipolar sensing seemed to be rather safe in magnetic fields with a flux density of up to 100 µT. Recently, Tiikkaja and coworkers36 investigated interference thresholds of cardiac pacemakers and ICDs at extremely low-frequency EMFs, but only in a small number of volunteers (13 ICD patients, 11 cardiac pacemaker patients) at magnetic flux densities not higher than 300 µT and not considering combined magnetic and electric exposure. None of the previous studies considered worst-case conditions, for example, maximum sensitivity of devices or full inspiration (see Test Procedure).
The present study investigated EMI with ICDs in a large in vivo study under worst-case conditions. We determined the lowest interference thresholds of 110 ICD patients in single and combined 50-Hz EMFs of up to 30 kV·m−1 and 2.55 mT.
The determined thresholds also apply for 60 Hz, the power frequency in the Americas. Previous studies showed that the susceptibility to EMI of cardiac implants is in the same range at 50/60 Hz.20,38 The limit values of the American National Standards Institute do not differ between 50 and 60 Hz (Table 1).
The knowledge of these interference thresholds closes the gap in the current guidelines for limiting exposure of ICD wearers to EMFs. Our data provides evidence that ICD disturbances do not occur within the limits values of the European Union for the general public (5 kV·m−1/0.1 mT). The first sensing failures were detected at stronger fields, and only 1 device could be disturbed in the atrial channel within the range of current European Union limit values for occupational exposure (10 kV·m−1/0.5 mT). However, new occupational guidelines are currently being discussed in the European Union parliament (up to 10 kV·m−1/6 mT at 50 Hz).8 Our data indicate that, should these limits come into action, ICD disturbances are more likely to occur.
Right Atrial Disturbances
In 19 of 31 patients (61.3%), an atrial oversensing was registered in EMFs within the tested limits (30 kV·m−1/2.55 mT). The higher probability of EMF interference in the atrial channel can be ascribed to the small intrinsic atrial signals with consecutively higher programmed atrial sensitivities and a corresponding poor signal-to-noise ratio in the atrium channel. Right atrial disturbances may lead to a scenario that potentially carries risks for patients; sustained oversensing can cause an inadequate mode switch to VVI(-R)/DDI(-R). If patients are in sinus rhythm, the asynchronous pacing mode may increase the risk for developing atrial fibrillation or pacemaker syndrome. The latter is caused by an atrioventricular dyssynchrony with subsequent loss of atrial contribution to ventricular diastolic filling and nonphysiological pressure waves. In case of chronotropic incompetence in patients with sick sinus node with an implanted 2- or 3-chamber device, a mode switch from DDD-R to VVI/DDI without activation of rate response may lead to the loss of chronotropic competence. In case of a dual-chamber device and sinus bradycardia, a pacing-induced left bundle-branch block pattern of activation with subsequent mechanical dyssynchrony can potentially lead to a loss of physical capacity. In the case of isolated atrial oversensing, a spontaneous VT may be inadequately classified as supraventricular tachycardia. If a supraventricular tachycardia time-out is not programmed, no therapy would be delivered in case of VT.
Right Ventricular Disturbances
Ventricular interferences caused by the field exposure were detected 20 times at maximum sensitivity and 5 times at nominal sensitivity. Right ventricular disturbances may lead to an inadequate detection of VT/VF and subsequent antitachycardia pacing or shock delivery. There is increasing evidence that inadequate shocks by themselves are associated with worse prognosis, although a clear cause-effect relationship has not yet been proven.39 Some devices interpret these signals, depending on the individual sensing algorithm, as an artificial noise (eg, short intervals <120–130 milliseconds are unlikely to be VF) and subsequently switch to a certain disturbance mode, which may be programmed at V00/D00/000. However, in this case, a spontaneous VT/VF episode cannot be detected. In cases of premature ventricular contractions or intrinsic heart rates higher than the programmed pacing rate of V00/D00, stimulation may lead to delivery of stimuli into the T wave, carrying the risk of VT/VF induction. This is not unlikely because strong EMFs occur mostly in occupational environments when the patient may be under physical stress and thus have an increased intrinsic heart rate and an overall higher likelihood of VT/VF occurrence (eg, because of ischemia in coronary artery disease) and elevated serum catecholamine levels.
Shorter or pulsed noise episodes may not trigger VT/VF detection and subsequent ICD therapy but may lead to pacing inhibition in patients with pacemaker dependency, which may cause symptomatic bradycardia or loss of resynchronization efficacy in CRT patients.
In Tables I and II in the online-only Data Supplement, worst-case conditions are shown for all interference thresholds determined.
In terms of the atrial channel, sustained pacing affected the interference thresholds in 15 of 41 runs (36.6%) under single or combined exposure. At full inspiration, the thresholds decreased in 5 of 41 runs (12.2%). Both conditions had an impact in 5 of 41 cases (12.2%).
For the ventricular channel, interference thresholds changed in 30 of 33 exposures (90.9%) during pacing. Respiration influenced the interference thresholds in only 1 patient (P095); sustained pacing had no impact. Thus, the results support the assumption of the influence of the worst-case conditions of full inspiration and sustained pacing.
Our results further confirm previously obtained data from pacemakers21 showing the influence of the programmable sensitivity on interference thresholds of ICDs. The susceptibility to EMI was coupled with the sensitivity settings, that is, the lower the sensitivity value, the lower the interference threshold of the ICD and vice versa (in the same patient). However, different patients with equal interference thresholds do not necessarily have the same sensitivity settings (Tables I and II in the online-only Data Supplement). The impact of the sensitivity values varies among manufacturers because of the manufacturer’s specific automatic adjustment of the sensitivity threshold. For example, when the ventricular disturbances of the Biotronik and St. Jude Medical single-chamber ICDs were compared, the data revealed that 4 of 26 tested Biotronik ICDs (15.4%) and 4 of 34 tested St. Jude Medical ICDs (11.8%) could be disturbed at sensitivity values of 0.5 and 0.2 mV, respectively.
The susceptibility to EMI of an ICD is also influenced by the type of lead and the patient’s physique.18,21 Table III in the online-only Data Supplement gives details of the leads and patient physique for all patients.
Potential Clinical Implications and Clinical Management
It is important to identify patients at risk of EMI. If strong EMF exposure is expected, particular care must be taken to optimize the implantation procedure (maximum achievable P/R potential). An ICD test with low-sensitivity settings has to be considered. Moreover, regular control of the intrinsic signal amplitudes (P/R wave) and the occurrence of EMIs via telemedicine transmitter is advisable.
It is not possible to define general sensitivity settings for EMF protection because of several individual factors, including lead position, patient physique, and type of EMF source. It remains a challenge for physicians to find a sensitivity level that gives a good balance between reduced EMI and accurate VT/VF sensing.
When inappropriate ICD discharges or episodes of EMI occur, patients should be assessed carefully. The situation of EMI occurrence should clearly be evaluated. Sometimes onsite measurements of EMFs are necessary. In terms of minimal device sensitivity, adjustment should be combined with ICD testing. Furthermore, patients should be tested in simulated EMFs, as in the present study.
In case of suddenly perceiving interference, increasing the distance to sources of EMF is the first remedial action to stop the dysfunction. Device defects caused by low-frequency EMFs have not yet been documented.
The present study was not designed to classify specific ICD models concerning their susceptibility in EMF exposure situations. However, further investigations may identify patient-, device-, and lead-related predictors of EMI.
Dual-chamber ICDs and CRT-Ds are underrepresented in this study. Therefore, conclusions on atrial interferences are not based on as many patients as for the right ventricular lead. In addition, the uneven distribution of the number of implants from each manufacturer may have influenced the results.
The validation of the electric field generation is based on a method comparison with 6 volunteers (see the online-only Data Supplement for details). Although the results indicate good agreement between the 2 methods, the data should be validated with a larger number of volunteers.
Furthermore, the findings of this study cannot be transferred to EMI at intermediate frequency and radiofrequency. Finally, the data are not applicable to pacemakers because of the difference in signal analysis of pacemakers and ICDs. Further study focusing on pacemaker patients is necessary.
The findings indicate that extremely low-frequency EMFs of everyday life do not disturb sensing capabilities of ICDs. The limit values for the protection of humans exposed to EMFs in general public assume to protect patients with ICDs at 50/60 Hz. In contrast, strong electric, magnetic, or combined fields in certain occupational environments are capable of causing undersensing or inappropriate sensing of atrial/ventricular tachyarrhythmias. However, a correct device function can still be expected in most cases. ICD devices with atrial sensing are more susceptible to EMI than single-chamber systems. Pacing in the ventricle increases the susceptibility to EMI.
In case of uncertainty about EMI, in vivo provocation examination such as those described in this study can provide a reliable and individual risk assessment for patients with implanted devices.
We thank the volunteers who participated in this study and the EMF-Portal team for the valuable contribution on the current status of publications on this topic and their editorial input to this manuscript.
Source of Funding
This study was funded through a grant from the German Social Accident Insurance Institution for the energy, textile, electrical, and media products sectors (BG ETEM) and the research unit for electropathology (FFE).
Drs Napp and Zink received travel grants from Biotronik, Boston Scientific, Medtronic, and St. Jude Medical. Drs Knackstedt, Bellmann, Marx, and Schauerte have received funding from Biotronik, Boston Scientific, Medtronic, and St. Jude Medical for consulting and lectures. The other authors report no conflicts.
To date, reliable systematic data on electromagnetic interferences on implantable cardioverter-defibrillators are scarce despite a high potential clinical relevance. Current recommendations by the manufacturers are very conservative with respect to exposure of implantable cardioverter-defibrillator patients to electric and magnetic fields (EMFs). This is based on the assumption that EMFs may lead to harmful interferences with the device. Recommendations on the code of behavior on how to handle electric and magnetic field sources in everyday life are inconsistent and are not based on in vivo studies. Decision making for implantable cardioverter-defibrillator implantation for the primary prevention in job-related EMF-exposed patients and subsequent recommendation of early retirement is often complex. National and international guidelines for the protection of humans exposed to EMF exclude patients wearing electric cardiac implants. The present study shows that electromagnetic interferences occur predominantly in relatively strong EMFs, which are normally present only in occupational environments. Moreover, we demonstrate a strong dependency on the programmed sensitivity of the device. Additionally, our data suggest that individual thresholds of electromagnetic interferences can be obtained and compared with the individual exposure of the patient. These results are important for clinicians to optimize the implantation procedure to achieve maximum obtainable intracardiac electrogram potentials, to choose appropriate device programming, and to provide advice for the management of patients with foreseeable high EMF exposure. Nonetheless, further investigations are needed to investigate patient- and device-related predictors of electromagnetic interferences. This may help to develop better sensing algorithms and to design new implantable cardioverter-defibrillator leads for the prevention of harmful electromagnetic interferences of implantable cardioverter-defibrillators.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.113.003081-/DC1.
- Received April 7, 2013.
- Accepted October 10, 2013.
- © 2013 American Heart Association, Inc.
- Brugada J,
- Vardas P,
- Wolpert C
- Auricchio A,
- Kuck KH,
- Hatala R,
- Arribas F
- 4.↵Institute for Applied Quality Improvement and Research in Health Care GmbH (AQUA). German Hospital Quality Report 2011. http://www.sqg.de/quality-report/index.html. 2012;53–58. Accessed April 4, 2013.
- Daubert JP,
- Zareba W,
- Cannom DS,
- McNitt S,
- Rosero SZ,
- Wang P,
- Schuger C,
- Steinberg JS,
- Higgins SL,
- Wilber DJ,
- Klein H,
- Andrews ML,
- Hall WJ,
- Moss AJ
- 6.↵Council of the European Union (CONSILIUM). Council recommendation of 12 July 1999 on the limitation of exposure of the general public to electromagnetic fields (0 Hz to 300G Hz). Official Journal of the European Communities. Brussels, Belgium;1999. 1999/519/EC.
- 7.↵Council of the European Union (CONSILIUM). Corrigendum to Directive 2004/40/EC of the European Parliament and of the Council of 29 April 2004 on the minimum health and safety requirements regarding the exposure of workers to the risks arising from physical agents (electromagnetic fields). Official Journal of the European Communities. Brussels, Belgium; 2004. 2004/40/EC.
- 8.↵Council of the European Union (CONSILIUM). Proposal for a Directive of the European Parliament and of the Council on the minimum health and safety requirements regarding the exposure of workers to the risks arising from physical agents (electromagnetic fields). Official Journal of the European Communities. Brussels, Belgium;2012. 14020/12.
- 9.↵IEEE Standards Association (IEEE-SA). IEEE standard for safety levels with respect to human exposure to electromagnetic fields, 0-3k Hz. New York, NY: The Institute of Electrical and Electronics Engineers;2002. C95.6-2002.
- 11.↵European Committee for Electrotechnical Standardization (CENELEC). Active implantable medical devices - Part 2-2: Particular requirements for active implantable medical devices intended to treat tachyarrhythmia (includes implantable defibrillators). Central Secretariat. Brussels, Belgium; 2008. EN45502-2-2:2008.
- 12.↵International Organization for Standardization (ISO). Implants for surgery: active implantable medical devices, part 6: particular requirements for active implantable medical devices intended to treat tachyarrhythmia (including implantable defibrillators). ISO Standards Catalogue. Geneva, Switzerland; 2010. ISO14708-6:2010.
- 13.↵International Organization for Standardization (ISO). Active implantable medical devices: electromagnetic compatibility: EMC test protocols for implantable cardiac pacemakers, implantable cardioverter defibrillators and cardiac resynchronization devices. ISO Standards Catalogue. Geneva, Switzerland; 2012. ISO14117-2012.
- 14.↵American Nationals Standards Institute/Association for the Advancement of Medical Instrumentation (ANSI/AAMI). Active implantable medical devices - electromagnetic compatibility - EMC test protocols for implantable cardiac pacemakers and implantable cardioverter defibrillators. Arlington;Association for the Advancement of Medical Instrumentation;2007. ANSI/AAMI PC69:2007.
- 15.↵Berufsgenossenschaft Energie Textil Elektro Medienerzeugnisse (BG ETEM). Unfallverhütungsvorschrift Elektromagnetische Felder. Hauptverwaltung. Koeln; 2002. BGV B11-6.01.
- 16.↵Deutsche Elektrotechnische Kommission im DIN und VDE. Betrieb von elektrischen Anlagen: Teil 100: Allgemeine Festlegungen. VDE-Verlag. Berlin, Germany;2009. DIN VDE 0105-100:2009.
- Bolz T,
- Bahr A,
- Gustrau F,
- Eichhorn KF,
- Hille S,
- Hentschel K
- Rauwolf T,
- Guenther M,
- Hass N,
- Schnabel A,
- Bock M,
- Braun MU,
- Strasser RH
- Roedig JJ,
- Shah J,
- Elayi CS,
- Miller CS
- Tiikkaja M,
- Aro AL,
- Alanko T,
- Lindholm H,
- Sistonen H,
- Hartikainen JE,
- Toivonen L,
- Juutilainen J,
- Hietanen M