Digital Cellular Telephone Interaction With Implantable Cardioverter-Defibrillators

1.1.1 Patients and Device Configuration

The study group consisted of 41 patients (7 women and 34 men). Twenty-one patients had a model 7217 PCD cardioverter-defibrillator (Medtronic, Inc.) implanted in an abdominal location, and 20 had a model 7219 Jewel ICD—18 in the pectoral region (6 subcutaneous and 12 deep to the pectoralis major muscle) and 2 in the abdominal region. Forty of the ICDs were connected to bipolar lead systems. One lead system consisted of an integrated bipolar design. Thirty-five of the patients had an implanted transvenous ventricular lead with 34 leads (Medtronic model 6936) consisting of polyurethane insulation and coaxial wound conductors, a high voltage coil, a ring and active fixation helix at 12-mm spacing for bipolar sensing and one lead (Cardiac Pacemaker, Inc., model 0115) consisting of silicone insulation with multilumen conductors, a high voltage coil and a passive fixation electrode with 12-mm spacing for integrated sensing implanted in the right ventricular apex. Six of the patients had an epicardial lead system consisting of two unipolar active fixation leads (Medtronic model 6917) with a nominal separation of 2 cm between electrodes, connected in a bipolar configuration. The separation between the lead conductors varied during the length of the lead, and any excess length was coiled at the implantation site, which would have very little effect on the antenna properties of the lead.

1.1.2 Digital Phones

The hand-held mobile telephone used for all tests was an AT&T model 6650 capable of analog and digital operation with 0.6 W peak power. The digital transmission mode in this North American Digital Communications/Time Division Multiple Access–50-Hz (NADC/TDMA-50) technology consists of an 828-MHz carrier signal modulated at a 20-ms pulse period (50 Hz pulse rate). This telephone had the advantage of displaying the actual power that was used during a call.

1.1.3 Study Protocol

All tests were thoroughly explained to each patient, and consent was obtained and coordinated with the primary care physician before commencing the test. All patients tested used the same mobile telephone at the same location in the clinic. The transmitting power of the mobile telephone was recorded for each patient test. All patients were connected to an ECG monitor and continuously observed during the tests. The tests were conducted with the telephone ringing (receiving) and transmitting (calling) 1) at both the right and left ears, simulating regular use; 2) moving over the entire route of the lead from electrode in the heart to its connector end; and 3) on top of the implantation site to simulate carrying the telephone in the upper shirt pocket or at the belt. All tests were with telephone configurations consisting of the antenna withdrawn (in) into the base (strongest magnetic field) and the antenna fully extended (out) using the tip (strongest electrical field). The ICDs tested in this study were able to provide noninvasive, telemetered, internal event counter information on detected episodes of ventricular tachycardia or ventricular fibrillation. Specific variables were reprogrammed to “worst case” values for this test (Table 1) to enhance the probability that detection of EMI from the cellular telephone would produce inappropriate device operation. The sensitivity was reprogrammed to the most sensitive value, and the number of intervals to detect tachycardia or fibrillation was set to the minimal value to satisfy the detection algorithm. All tachycardia pacing and shock therapies were temporarily disabled as a precaution in the event that the EMI was sensed, thereby initiating therapy. The pacing rate was increased ∼10 ppm above the patient’s intrinsic rate to produce continuous overdrive pacing. The programmer head was not left in place over the ICD during testing.

The defibrillator was interrogated after the patient completed each specific telephone test to determine whether the ventricular tachycardia or fibrillation detection algorithms were satisfied by detection of EMI. The telemetered data would identify any detected episode of ventricular tachycardia or fibrillation by indicating an increment in the internal counter used to monitor for arrhythmia detection in the Tachycardia Counter Data Report. All variables were reprogrammed to original values after completion of the test.

The cellular telephone has a built-in speaker with a small, permanent magnet in the earpiece that generates a static magnetic field. A magnetic field could activate a reed switch in the cardioverter-defibrillator and temporarily inhibit detection of ventricular tachycardia or fibrillation. However, this magnetic field interference would not be detected in this in vivo study, because, unlike standard pacemakers, the pacing circuit of the ICDs being tested will not revert to an asynchronous pacing mode. Therefore, this test could not be done in vivo, but was accomplished with a custom device during in vitro testing (see In Vitro Testing, Magnet test).

1.1.4 Statistical Analysis

An exact binomial upper 95% confidence limit was used on the failure rate.

1.2.1 Antenna-Generated Field

The ICD was connected to either a transvenous or epicardial lead system. The complete system was then immersed in a 375 ± 15 ohm cm saline solution bath and placed within 1 cm (typical implant depth) of the liquid surface in a frontal orientation, as shown in Fig. 1. A transmitting antenna was connected to a frequency pulse generator to simulate cellular telephone modulation. The testing was structured to assess susceptibility of the ICD’s normal operation to modulations that are known to be used in two digital cellular telephone systems: NADC/TDMA-50 and international Global System Mobile (GSM). Four models of ICDs—PCD 7217, Jewel 7219, Jewel Plus 7220 and Micro Jewel 7221—were subjected to radiated continuous carrier wave and modulated interference fields at two frequencies—836 MHz and 900 MHz. Modulations that simulate the digital systems patterns as shown in Table 2. The field intensity was increased until the ICD began to sense the interference, and the level at that point was recorded.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1

Top view of tank simulation model used for the antenna-generated field in vitro test of the ICDs. Amp = amplifier; Atten = attenuator; EFG-2 = electric field generator-2; Gen. = generator; RF VM = radiofrequency volt meter.

1.2.2 Analog/Digital Cellular Phones

The ICDs listed in Table 3immersed in saline, were subjected to the near fields radiated from the antennas of the cellular telephone models used in North America and worldwide, as shown in Table 4. The telephones were tested in both analog and digital modes, where applicable. Some telephones, designated as group 1 in Table 4could be fully operated, and the susceptibility was evaluated while the telephone was transmitting and receiving calls with the antenna withdrawn (in) and extended (out). Other technologies, designated as group 2 in Table 4were used in an open-loop transmit mode (i.e., producing the typical pulsing format of that technology without being in communication with an active cell site or base station simulator). Keypad programming was used to configure the telephone in the full power transmit mode. The distance from the telephone antenna to the ICD was kept within 1 cm.

1.2.3 Magnet Test

A custom ICD was modified with special software so that when the reed switch was activated with a magnet, the pacing rate would change to a train of 7 pulses of low amplitude and a fast rate (100 ppm), and then return to normal pacing. The cellular telephone, with its magnet in the earpiece, was brought into direct contact with the ICD in free air and then moved around on and at incremental distances from the surface while the pacing output was monitored.

2.1.1 Digital Cellular Phone

The ICD was interrogated after each phase of telephone status and location testing. At no time was any ventricular tachycardia or fibrillation counter activated by the EMI from the antenna of the cellular telephone for any test conducted on any patient. There was no damage or reprogramming of any ICD during the tests. The ECG that was recorded continuously during the patient test to monitor the ventricular demand overdrive pacing mode (VVI) was thoroughly reviewed for abnormalities. At no time were there any indications of pacing inhibition or changes to the programmed pacing intervals during any test on any patient. Therefore, the binomial upper 95% confidence limit for the failure rate of 0% is 7%.

2.2.1 Antenna-Generated Field Testing

Field strength levels at which the ICDs began to sense the interference started at 230 V/m root-mean-square (RMS). Interference oversensing resulted in partial inhibition (changed pacing rates); satisfied the ventricular tachycardia or ventricular fibrillation detection algorithm and could have caused either therapy to be delivered; or caused the device to go power-on-reset. The most susceptible ICD to EMI was the early generation model 7217 connected to a bipolar epicardial lead system that occurred at field intensities of 230 V/m RMS. A bipolar, coaxial, polyurethane, insulated, endocardial lead connected to the model 7217 ICD increased the susceptibility interference level to 300 V/m RMS.

2.2.2 Predicted Field Intensity

To estimate the field values in the vicinity of the ICD, the predicted electric field intensity for a 0.6-W isotropic antenna point source, as recorded at measured distances in free space from the antenna, is plotted in Fig. 2. The predicted electric field intensity (E) for a 0.6-W isotropic antenna point source, as measured along axes in all directions for free space conditions (air at 377 Ω), was calculated at incremental distances from the transmitting antenna using the following equation [6]: E=(30P)1/2R(volts/meter),where P = transmitted power in watts (0.6 W), and R = distance in meters.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2

Theoretically predicted electric field intensity for a 0.6-W isotropic point source antenna, as measured in axis in all directions in air at incremental distances from the antenna.

The field intensity decreases exponentially and rapidly drops <50 V/m RMS at 10 cm from the antenna. At 15 cm (6 in.), the distance recommended for cellular telephone use for patients with ICDs, the field intensity would be 20 V/m RMS, which is well below the susceptibility level measured during the in vitro antenna-generated field testing. Field intensity predictions at distances <2 cm from the point source were not attempted, as this near field also contains magnetic fields that affect the accuracy of the equation. The near field characteristics for the cellular technologies were assessed by in vitro testing with the analog and digital cellular telephones.

2.2.3 Analog/Digital Cellular Telephones

Testing was completed and the results summarized (Table 3) on models of market-released ICDs, and testing was completed on various lead systems, with the analog/digital cellular telephones in groups 1 and 2 listed in Table 4. No interference was noticed on any ICD with any cellular telephone technology.

2.2.4 Magnet Test

The static magnetic field generated by the magnet of the speaker in the telephone was measured to be 288 gauss RMS at the telephone surface and decreases exponentially to 19 gauss at 1 cm (0.4 in.) from the surface. The cellular telephone placed over and moved around on the ICD surface activated the reed switch when the telephone was on the upper left corner of the engraved side, where the reed switch is located. When the cellular telephone was placed at 0.5 cm (0.2 in.) or further from the ICD surface, reed switch activation did not occur.