Low risk of electromagnetic interference between smartphones and contemporary implantable cardioverter defibrillators

Transcription

1 Europace (2016) 18, doi: /europace/euv374 CLINICAL RESEARCH Sudden death and ICDs Low risk of electromagnetic interference between smartphones and contemporary implantable cardioverter defibrillators Haran Burri*, Louis Paulin Mondouagne Engkolo, Nicolas Dayal, Abdul Etemadi, Anne-Marie Makhlouf, Carine Stettler, and Florence Trentaz Cardiology Service, Geneva University Hospitals, Rue Gabrielle Perret-Gentil 4, Geneva 1205, Switzerland Received 7 September 2015; accepted after revision 15 October 2015; online publish-ahead-of-print 8 February 2016 Aims Manufacturers of implantable cardioverter defibrillators (ICDs) recommend that cell phones be maintained at a distance of 15 cm from the implanted device in order to avoid risk of dysfunction due to electromagnetic interference (EMI). Data relating to this issue are outdated and do not reflect modern technology. Our aim was to evaluate whether EMI is still an issue with contemporary ICDs and smartphones.... Methods Consecutive patients implanted with a wireless-enabled ICD were tested for potential interference with two models of and results recent 4G smartphones in conditions intended to maximize risk of EMI. A magnet effect (due to the phone speakers) was tested by placing the smartphones in the standby mode directly over the ICD generator. The presence of EMI artefacts on the real-time electrograms was evaluated by placing the smartphones in the standby, dialling, and operating modes directly over the generator casing and over the parasternal region in the vicinity of the ventricular lead. A total of 63 patients equipped with 29 different models of single, dual, or biventricular ICDs from five major manufacturers were included. None of the patients showed any evidence of interference with the smartphones during any of the 882 tests.... Conclusion The risk of EMI between modern smartphones and contemporary ICDs is low. This is probably due to the filters incorporated in the ICDs and low emission by the phones, as well as the small size of the magnets in the smartphones tested.... Study NCT ( registration Keywords Mobile phone Smartphone Implantable cardioverter defibrillator Electromagnetic interference Magnet effect Introduction The telecommunications industry has been the stage of remarkable technological advances in the last decade, notably within the domain of mobile phones. With the advent of third- (3G) and fourth (4G)-generation communication standards, their use has grown exponentially. According to the International Telecommunications Union ( the number of active mobile phone subscriptions in 2015 is estimated at 1.2 per inhabitant in the developed world. Mobile phones generate two different types of electromagnetic radiation (EMR) fields: a static field produced by the magnets located in the speakers (in the earpiece and loudspeaker), and a dynamic field resulting from the emission of radiofrequency necessary for communication. Numerous reports in the 1990s have shown electromagnetic interference (EMI) between cell phones and pacemakers, although this is less of an issue with modern devices. 1,2 Implantable cardioverter defibrillators (ICDs) potentially have a higher risk of interference than pacemakers due to the high programmed sensitivity. Electromagnetic radiation may cause oversensing with inhibition of pacing or inappropriate therapy, as well as temporary inactivation of antitachycardia therapy due to a magnet response. Previous in vitro and in vivo studies have shown EMI between mobile phones and ICDs. 3,4 Based on these results and * Corresponding author. Tel: ; fax: address: haran.burri@hcuge.ch Published on behalf of the European Society of Cardiology. All rights reserved. & The Author For permissions please journals.permissions@oup.com.

2 Electromagnetic interference between smartphones and contemporary ICDs 727 What s new? No cases of electromagnetic interference (EMI), either with signal artefacts or a magnet effect, were found between two popular models of smartphones and contemporary ICDs from five major device manufacturers. The risk of EMI seems to be currently negligible, and the recommendations made to patients to hold their cell phones at 15 cm from their implantable device should be revised in order to avoid unnecessary stress and allow them to lead lives that are as normal as possible. also on regulations for testing electromagnetic compatibility of implanted devices, 5 all manufacturers of cardiovascular implantable electronic devices (CIED) currently recommend a safety distance of 15 cm between mobile phones and the implanted device. However, these reports date from at least a decade and may not reflect current risk due to the presence of filters in more recent CIEDs as well as a lower emission power of more recent mobile phones. It should also be noted that all these studies were limited to previous generation cell phones operating on the GSM (2G) bandwidth of 900 MHz and most were carried out with the protective effect of the programmer head placed over the generator. Current smartphones use the UMTS (3G) standard (which has different signal properties compared with GSM) for phoning and LTE (4G) standard for data transfer (see Table 1). To the best of our knowledge, the risk of EMI of contemporary mobile phones and ICDs has not yet been published. Due to the growth of numbers of ICD patients exposed to cell phones, this issue has reached public health proportions. We therefore sought to evaluate the risk of EMI between contemporary smartphones and ICDs. Methods Patients and devices We enrolled consecutive outpatients scheduled for routine ICD followup at our centre. All types of ICDs were included in our study (single chamber, dual chamber, biventricular). The only inclusion criterion was the presence of a wireless-enabled ICD, as this allowed continuous communication between the programmer and the implanted device during all tests, without the necessity of maintaining the programming head on the generator (which may have a shielding effect, cause loss of telemetry, or generate trains of artefacts). 6,7 Patients who were pacemaker dependent or aged,18 years, as well as those unable to sign a consent form, were excluded from the study. The study was approved by the institutional ethics committee, and all patients provided written informed consent. Study protocol Two popular models of smartphones were used: the Samsung Galaxy S4 with an emission Specific Absorption Rate (SAR) of 0.5 W/kg and the Apple iphone 6 (SAR 0.9 W/kg), both operating on the same mobile network (Sunrise Switzerland). Both smartphones were quad band, operating at 850/900/1800/1900 MHz. The static magnetic field strengths of the cell phone speakers were measured using the magnetic sensor of a Samsung Galaxy SIII phone and the Sensor Kinetics Android App. In the standby mode with good network field strength, a mobile phone has negligible emission. The emission is considerably increased in case of low network with automatic switching of bands by the phone to search for a better signal. In order to increase the emission power of the mobile device (and hence the risk of EMI), all tests were conducted in a room with a weak mobile network located in the hospital basement adjacent to the electrophysiology lab. The smartphones displayed a maximum of 1 bar on the reception indicators, and the field strength was consistently measured as being less than 2100 dbm using an Android application on the Samsung device (KAIBITS Software v , Frankfurt, Germany). The presence of a magnet effect was evaluated by placing each smartphone in the standby mode over the ICD generator and noting the response shown in Table 2. Routine ICD interrogation was performed and included assessment of background noise on the electrograms (EGMs). All tachyarrhythmia therapies were then inactivated to avoid inappropriate therapy in case of oversensing of EMI, without changing pacing parameters. In order to increase the risk of interference, the sensitivity of the atrial and ventricular leads was set to the maximum programmable level (see Table 2). Evaluation of EMI on the EGMs consisted of placing the smartphone directly over the generator casing in the standby mode, during ringing and during operation of a call. The same sequence was repeated with the smartphone placed in the left parasternal region, at proximity to the right ventricular lead. During all the tests, the intracardiac EGMs were monitored and printed in real time to assess for visible artefacts (comparing the tracing with the Table 1 Overview of cell phone transmission technology (shown for current standards in Switzerland) Generation Voice Data Carrier band (MHz) Pulsing (Hz) Max. power (W)... 2G GSM EDGE G UMTS HSPA (900) 4G VoLTE LTE (900 Hz) 1 (1800Hz) GSM, global system for mobile communications; LTE, long-term evolution; EDGE, enhanced data rates for GSM; HSPA +, high-speed packet access plus; VoLTE, voice over LTE (only available in very recent devices).

3 728 H. Burri et al. Table 2 Maximum programmable sensitivity and magnet response according to device manufacturer Atrial (mv) Right ventricle (mv) Left ventricle (mv) Magnet response... Biotronik / Episode list a Boston Scientific Audible alert Medtronic N/A Audible alert St-Jude Medical N/A Vibratory alert Sorin N/A Pacing at 96 bpm Different models of the same manufacturer may have different sensitivity settings. N/A, not available. a Indication of magnet application on the episode list upon ICD interrogation after having placed the phone for at least 5 min on the device. recording prior to exposure to the smartphone to discount for background noise) or oversensing (by analysing the marker channels). The tests were performed consecutively with each of the two smartphones in all patients, with a total of 14 tests per patient (in addition to the baseline evaluation). At the end of the procedure, the ICD was reinterrogated and reprogrammed to the initial settings. Results A total of 63 patients were included, with a total of 882 tests (in addition to baseline evaluation of the ICDs). Patient characteristics are shown in Table 3, with the 29 different ICD models from 5 major manufacturers shown in Table 4. IS-4 connectors were present in 11 CRT-Ds and DF-4 connectors in 13 devices (Boston Scientific, Medtronic, and St-Jude Medical models) with DF-1 and IS-1 connectors in all other devices. All generators were implanted in a left prepectoral pocket, and respective leads were implanted in the right atrial appendage, right ventricular apex, and postero-lateral coronary sinus tributaries. The leads corresponded to the same manufacturers as the generator (with all Boston Scientific ICD leads being of integrated bipolar design). Sensing was programmed to bipolar in all leads, i.e. none of the left ventricular leads were programmed to extended bipolar sensing. For the Galaxy 4 phone, the magnetic field was measured at 1.5 mt at 0 cm and at 0.02 mt at 2 cm. For the iphone 6, the field was measured at 1.2 mt at 0 cm and 0.01 mt at 2 cm. There were no cases of EMI either detected by the ICDs or visualized on the real-time EGM recordings. Furthermore, there was no evidence of a magnet effect resulting from proximity of the smartphones in any of the patients studied. Table 3 Patient characteristics Patient characteristics (n 5 63)... Age (years) Male gender 54 Weight (kg) Height (cm) Heart rate (bpm) LVEF (%) Cardiopathy aetiology, n (%) Ischaemic 28 (44) Non-ischaemic 35 (56) NYHA class I 14 (22) II 33 (52) III 15 (24) IV 1 (2) ICD type, n (%) Single chamber 13 (21) Dual chamber 18 (29) CRT-D 32 (51) Device indication Primary 39 (62) Secondary 24 (38) Time since implantation (months) Data are expressed as mean + standard deviation. Percentages are shown in parentheses. LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; ICD, implantable cardioverter defibrillator; CRT-D, cardiac resynchronization therapy-defibrillator. Discussion The main finding of the study is that the smartphones did not cause any EMI in patients with contemporary ICDs, despite testing in the worst-case conditions that maximized the risk of interference. Our findings are in agreement with previous studies that showed lack of EMI between 2G GSM cell phones and ICDs All modern CIEDs have sense-amplifier filters that are centred on the frequency range of endocavitary signals generated by atrial and ventricular depolarization ( Hz). The high-pass filters aim to eliminate low-frequency signals (e.g. T-wave components), and the low-pass filters attenuate myopotentials and EMI. Although these filters may not eliminate 100% of the signal in the filtered bandwidths, and EMI with frequencies within the Hz spectrum will not be eliminated, these filters attenuate GSM signals that are of MHz frequencies. 11 However, GSM waveforms are modulated, which includes amplitude modulation and pulsing of the signal. An in vitro study by Barbaro et al. 11 has shown that the GSM signal may be demodulated by the pacemaker, probably by internal nonlinear elements of the circuit, yielding frequencies that may escape filtering and be sensed by the device (see Figure 1A). In addition to

4 Electromagnetic interference between smartphones and contemporary ICDs 729 Table 4 Implantable cardioverter defibrillator manufacturers and models Manufacturer (n) ICD type (n)... Medtronic (20) Evera XT VR (2); DR (1) Protecta DR (1); CRT-D (5) Secura VR (1); DR (1) Viva quad XT CRT-D (4) Viva XT CRT-D (5) Boston Scientific (15) Autogen X4 CRT-D (3) Cognis CRT-D (2) Incepta DR (2); CRT-D (2) Teligen VR (3); DR (2) Vitality2 VR (1) St-Jude (13) Ellipse VR (1); DR (2) Fortify ST DR DR (2) Quadra Assura CRT-D (4) Unify CRT-D (4) Biotronik (11) Ilesto 7 VR (1); DR (2); CRT-D (2) Iperia VR (2) Lumax VR (1); DR (2); CRT-D (1) Sorin (4) Paradym RF VR (1); DR (3) ICD, implantable cardioverter defibrillator; VR, single chamber; DR, dual chamber; CRT-D, biventricular defibrillator. the sense-amplifier bandpass filters, modern CIEDs are equipped with a second type of filter called feedthrough capacitor filters (see Figure 1B), which have been shown to be a major factor for reducing digital cell phone EMI with pacemakers (from 28.9 to 0.4% of patients, P ¼ 0.01). 12 These broadband filters (30 MHz 10GHz 13 ) avoid conduction of EMI from the CIED casing to the internal circuitry via the device header (which also occurs if the device is programmed with bipolar sensing). 11 The titanium case protects the device circuitry from EMI, but the interference may penetrate via the unshielded header, as shown by in vitro 4 and in vivo 3 studies with ICDs. These studies also showed that the lead tip and body are unlikely to conduct EMI. In fact, leads placed in a conductive medium (such as blood) are poor antennas. Furthermore, recent leads have nanomagnetic insulation to shield them from radiofrequency and magnetic fields. 14 Other factors that may contribute to reduced risk of EMI with modern cell phones are lower maximum power emission with the 3G and 4G standards ( 0.25 W compared with 1 2 W with 2G, and up to 8 W used in previous studies), 15 improved signal quality (and hence lower cell phone radiation) due to better provider coverage, less low-frequency modulation of the signal, as well as higher carrier band frequencies (resulting in less tissue penetration). Regulations in Europe and in the USA require that device manufacturers test for possible cell phone EMI in the 450 MHz-3 GHz bandwidth at 120 mw at 2.5 cm, which corresponds to emission of a phone placed 15 cm from the device. Voluntary testing may be performed at 8 W (at 450 MHz 1 GHz) and 2 W (at 1 3 GHz), which renders the implanted device compatible with cell phones without restrictions of distance. Despite the fact that this voluntary testing is usually performed, device manufacturers still maintain their warnings of keeping cell phones at a distance of 15 cm, in order to avoid possible legal issues. In a series presented in abstract format, of 161 ICD patients exposed to three smartphone models operating at maximum power output, a single case of EMI occurred in a patient with a Biotronik Lumax 640 CRT-D with two smartphones operating in GSM and UMTS modes. 16 However, two other patients in this series implanted with an identical model showed no EMI, and none of the 11 patients in our report equipped with a Biotronik ICD had any EMI. Reasons for these findings are unclear, but isolated cases of EMI despite feedthrough filters may occur if (1) the filter is defective (despite extensive testing by manufacturers), (2) leads are improperly connected to the header, and (3) the filter circuit is not properly in contact with the generator housing. These circumstances are likely to be exceptionally rare and should not affect recommendations made to patients. Another consideration is that CIED magnetic switches may be activated if a magnetic field is aligned with the switch at 1 mt (which is equivalent to 10G). 17 Cell phone speakers may be equipped with neodymium iron boron magnets, which are small in size but produce strong static magnetic fields (even if the phone is turned off). One in vitro study showed that these magnets may activate pacemaker and ICD magnetic switches. 18 However, the magnets in mobile phones are considerably smaller than those used in this study (which generated fields of at least 1.2 T). Another study reported interference with clip-on (but not in-ear) headphones in 38% of ICD patients, but only if the field strength of the headphone was 1 mt at a distance of 2 cm. 19 During in vitro testing, a magnet effect with ICD reed switch activation was found to be present when a cell phone with a magnetic field strength of 28.8 mt was placed,5 mm from the region of the casing overlying the reed switch, but did not occur during in vivo testing. 9 Our study did not show any magnet effect with newer-generation mobile phones in which the static magnetic field was measured,0.1 mt at 2 cm. Study limitations A limitation of our study is the relatively small number of patients, although the sample size is larger than all previous studies published in ICD patients. 3,6 9 The results of our study are applicable to the smartphones and ICDs tested, and we therefore cannot exclude interaction between other models of mobile phones and ICDs, although this is highly unlikely. Nevertheless, it would be of interest to test cell phone EMI with subcutaneous ICDs. None of our CRT-D patients had left ventricular sensing programmed to extended bipolar, but this is unlikely to have affected the results as cardiac leads have not been found to play a major role in EMI with cell phones. 3,11 We did not assess whether the smartphones interfere with transmission of remote monitoring data, but this has been previously shown to not be an issue. 20 We also did not evaluate whether the most recent voice transfer mode (4G VoLTE) and

5 730 H. Burri et al. A Modulated signal B Circuitry Demodulated signal Sense amplifier Amplified demodulated signal Bandpass filter Rectifier Level detector VF VF VF Sensitivity level Filtered and rectified signal Sensed EMI Ceramic feedthrough capacitor Sense amplifier Bandpass filter Level detector Rectifier Figure 1 Schematic representation of sensing of EMI between cell phones and ICDs. (A) Situation without ceramic feedthrough capacitors. Cell phone emission (GSM signal waveform with MHz range), which is demodulated by the ICD circuitry into a signal that includes low frequencies (in the Hz range). The signal is then amplified by the sense amplifier before being filtered (bandpass Hz), retaining only the low-frequency components, which are then sensed by the device. (B) The leads have an insignificant antenna effect, and EMR is reflected by the titanium generator casing. Electromagnetic radiation may, however, penetrate the device via the unprotected header and may then cause oversensing via the mechanism described in A. However, the addition of a ceramic feedthrough capacitor blocks the entry of EMR into the device circuitry and thereby avoids oversensing. Table 5 Simplified EMI testing procedure between cell phones and ICDs Testing for interference with EMR Inactivate ICD therapies. Evaluate the strength of the network field (the lower the field, the higher the phone output power) and the carrier band (3G, 4G, etc.) on the cell phone display. Program maximum sensitivity of all ICD sensing channels. Visually evaluate baseline artefacts and the presence of oversensing (e.g. of T-waves). Place the phone in the standby mode over the generator. Place an incoming call to the cell phone. Evaluate for the presence of EGM artefacts and oversensing (marker channel during wireless telemetry). Testing for a magnet effect Place the cell phone in the standby mode over the ICD generator in a vertical axis (as when placed in a pocket). Optionally, the phone may also be rotated over 1808 to test for different magnetic field alignments. Move the cell phone over the ICD generator (to scan proximity with the reed switch or Hall sensor). Evaluate the presence of a magnet response according to the manufacturer (cf. Table 2). data transfer modes (EDGE, HSPA+, and LTE) result in EMI, but this is highly unlikely because these modes either have similar characteristics as the voice modes (for EDGE and HSPA+) orhavelow power and the absence of low-frequency signal components (for HSPA+, VoLTE,andLTE).Likewise,BluetoothandNearField Communication (NFC) were not tested for EMI but are also unlikely to have affected results because of the very low power ( 1 mw maximum power for Bluetooth and 0.1 W for NFC). Conclusion Our study demonstrates the absence of interference between contemporary smartphones and ICDs, which is in line with data published over the last decade. Due to technological advances, both in mobile phone and ICD technology, the risk of EMI seems to be currently negligible. It is likely to remain so for future generations of telecommunication standards duetolowemissionpowerof

6 Electromagnetic interference between smartphones and contemporary ICDs 731 the phones, high signal frequencies (with reduced tissue penetration), and effective CIED filters. As such, this study provides important safety information, but ideally, individual testing of patients for lack of interaction with their cell phone should be performed at device follow-up (a simple testing protocol is shown in Table 5). However, it should be borne in mind that the device clinic may not represent a worst-case scenario due to a good network field strength. Lifestyle recommendations issued to patients should be evidence based, pragmatic (and accept the fact that risk may not be zero), and should not cause unnecessary stress. If testing does not show any EMI, patients may be reassured regarding use of their phone, which should help them lead lives that are as normal as possible. Acknowledgements The authors wish to thank Dr Hugo Lehmann (Swisscom), Mr Guillaume Girard (Medtronic), and Mr Thomas Dörr (Biotronk) for their replies to our technical queries and review of the manuscript. Conflict of interest: none declared. References 1. Censi F, Calcagnini G, Triventi M, Mattei E, Bartolini P. Interference between mobile phones and pacemakers: a look inside. Ann Ist Super Sanita 2007;43: Hekmat K, Salemink B, Lauterbach G, Schwinger RH, Südkamp M, Weber HJ et al. Interference by cellular phones with permanent implanted pacemakers: an update. Europace 2004;6: Barbaro V, Bartolini P, Bellocci F, Caruso F, Donato A, Gabrielli D et al. Electromagnetic interference of digital and analog cellular telephones with implantable cardioverter defibrillators: in vitro and in vivo studies. Pacing Clin Electrophysiol 1999;22: Bassen HI, Moore HJ, Ruggera PS. Cellular phone interference testing of implantable cardiac defibrillators in vitro. Pacing Clin Electrophysiol 1998;21: Calcagnini G, Censi F, Bartolini P. Electromagnetic immunity of medical devices: the European regulatory framework. Ann Ist Super Sanita 2007;43: Chiladakis JA, Davlouros P, Agelopoulos G, Manolis AS. In-vivo testing of digital cellular telephones in patients with implantable cardioverter-defibrillators. Eur Heart J 2001;22: Occhetta E, Plebani L, Bortnik M, Sacchetti G, Trevi G. Implantable cardioverter defibrillators and cellular telephones: is there any interference? Pacing Clin Electrophysiol 1999;22: Sanmartin M, Fernandez Lozano I, Marquez J, Antorrena I, Bautista A, Silva L et al. The absence of interference between GSM mobile telephones and implantable defibrillators: an in-vivo study. Groupe Systemes Mobiles. Rev Esp Cardiol 1997; 50: Fetter JG, Ivans V, Benditt DG, Collins J. Digital cellular telephone interaction with implantable cardioverter-defibrillators. J Am Col Cardiol 1998;31: Tandogan I, Ozin B, Bozbas H, Turhan S, Ozdemir R, Yetkin E et al. Effects of mobile telephones on the function of implantable cardioverter defibrillators. Ann Noninvasive Electrocardiol 2005;10: Barbaro V, Bartolini P, Calcagnini G, Censi F, Beard B, Ruggera P et al. On the mechanisms of interference between mobile phones and pacemakers: parasitic demodulation of GSM signal by the sensing amplifier. Phys Med Biol 2003;48: Hayes DL, Wang PJ, Reynolds DW, Estes M 3rd, Griffith JL, Steffens RA et al. Interference with cardiac pacemakers by cellular telephones. N Engl J Med 1997; 336: Stevenson RA. Design and application of broadband ceramic feedthrough capacitor EMI filters to cardiac pacemakers and implantable defibrillators. Proc 19th Ann Int Conf IEEE 1997;6: Beinart R, Nazarian S. Effects of external electrical and magnetic fields on pacemakers and defibrillators: from engineering principles to clinical practice. Circulation 2013;128: Naegeli B, Osswald S, Deola M, Burkart F. Intermittent pacemaker dysfunction caused by digital mobile telephones. J Am Col Cardiol 1996;27: Lennerz C, Pavaci H, Grebmer C, Vrazic H, Semmler V, Kottmaier M et al. Electromagnetic interference between smartphones and current cardiac implantable electronic devices: rare but present (abstract). Europace 2015;S3: iii Jacob S, Panaich SS, Maheshwari R, Haddad JW, Padanilam BJ, John SK. Clinical applications of magnets on cardiac rhythm management devices. Europace 2011; 13: Wolber T, Ryf S, Binggeli C, Holzmeister J, Brunckhorst C, Luechinger R et al. Potential interference of small neodymium magnets with cardiac pacemakers and implantable cardioverter-defibrillators. Heart Rhythm 2007;4: Lee S, Fu K, Kohno T, Ransford B, Maisel WH. Clinically significant magnetic interference of implanted cardiac devices by portable headphones. Heart Rhythm 2009;6: Calcagnini G, Censi F, Floris M, Pignalberi C, Ricci R, Biancalana G et al. Evaluation of electromagnetic interference of GSM mobile phones with pacemakers featuring remote monitoring functions. Pacing Clin Electrophysiol 2006;29: