Risks to pacemaker patients undergoing magnetic resonance imaging examinations

This editorial refers to ‘Low-field magnetic resonance imaging: increased safety for pacemaker patients?’ by K. Strach et al. , on page 952 Strach et al.1 report on successful magnetic resonance imaging (MRI) examinations in 114 patients using an open-scanner design with a magnetic field strength of 0.2 T. They list 12 items as potentially adverse effects: Before critically commenting on these real or alleged risks, an observation made in the midst of the 1980s with Medtronic pulse generators (PGs) of the Spectrax family is reported. There were case reports that these PGs developed several malfunctions during RF electrosurgery: reduction in output pulse, switching to back-up mode, inhibition, or total loss of function. Applying 500 kHz voltages to the input of a Spectrax PG, there was obviously rectification of the RF voltage inside the PG diminishing the battery voltage and thereby the output pulse, and causing the PG to enter the back-up mode. With still higher voltages, complete inhibition took place due to activation of the ‘run-away protection’. Reduction of the battery voltage to 1 V or less forced the circuitry into instability where the logic elements were neither ‘1’ nor ‘0’, resulting in high current consumption. If the battery resistance was 10 kΩ or higher, this state was no longer reversible when the interference [electromagnetic interference (EMI)] was ceased, thus yielding complete exhaustion within a short time. These phenomena no longer exist in modern Medtronic PGs. movement of pacemaker device or leads, radiofrequency (RF)-induced heating of lead tip, pacemaker dysfunction (induction of ventricular fibrillation and risk of asynchronous pacing), damage, rapid pacing, reed switch malfunction with or without asynchronous pacing, inhibition of pacing output, change or loss of programmed data caused by electrical reset, depletion of the pacemaker battery voltage, change in pacing capture threshold, high abandoned lead risk, and back-up mode creating cardiac insufficiency in children due to reduction in rate or output. Another phenomenon, but not connected to Medtronic PGs, arose when the PG was protected against external defibrillation by a simple Zener diode. The often used Z 7 limited the EMI at the input of the PG to 7 V (Zener voltage) in one polarity and to 0.7 V (diode voltage) in the other polarity. All voltages higher than the diode or Zener voltages were short-circuited by the Zener diode, so that stronger currents flowed through the electrode/tissue interface. Fibrillation could result from these rectified currents. This effect of pacemaker EMI was well known from investigations of RF electrosurgery and pacemakers. 2 Orland and Jones 3 even proved that fibrillation was possible if the PG was exhausted. Modern PGs are now protected by back-to-back Zener diodes or similar circuits that avoid rectification. There is no doubt that the asymmetric Zener diode protection will also short-circuit RF voltages during MRI, yielding scan synchronous stimulation pulses as described by several authors. 4–7 Against this background of RF experience, pacemaker experts engaged in the problem of MRI and pacemaker compatibility were at that time focused on what could possibly happen. Their experimental experience of whether these assumptions were realistic was limited in the 1980s. This should give rise to the question as to which of the adverse effects are realistic and which are without any experimental or clinical proof. To mention the latter is only of historical interest and should not be described as a threat that still exists. One phenomenon that was not discussed in the paper by Strach et al.1 or in others either 8 is the EMI behaviour due to gradient fields. Assuming that the asynchronous magnet mode is switched off or that the reed switch remains open inside the MRI machine, the PG will switch to its ‘noise mode’ if it is exposed to EMI from the strong voltages pulses—we measured in vitro 180 mV 9 —induced by gradient fields in unipolar or bipolar leads, whether atrial or ventricular. As these signals have a repetition rate of 8 Hz or more, all PGs will operate in their asynchronous ‘noise mode’. 9 In noise mode PGs stimulate asynchronously at their programmed basic rate. We found in 107 DDD PGs that all of them switched to DOO if both channels were exposed to EMI. 10 This means that EMI cannot force a dual-channel system to stimulate at the upper rate limit. Corresponding reports in the literature are probably a misinterpretation when the scan period or multiples of it are close to the upper rate period. Asynchronous pacing in the noise mode has the big advantage over magnet mode that the rate can be chosen individually and it can be stopped immediately by switching off the gradient fields. Inhibition cannot take place except when scanning starts (one missed beat) or when a scan period of 150 ms or shorter is followed by a pause of at least another 150 ms. 9 This pattern would be regarded by the PG as a heart rhythm that would inhibit it. Such interval scanning may occur in auditory examinations and are, thus, contraindicated in pacemaker-dependent patients unless pacemaker or ECG-triggered scanning is used. 9 Strach et al. claim that abandoned leads possess a higher risk than leads in use. 1 Between them, there is little physical difference, in that a lead connected to a PG can be regarded as being short-circuited by the input capacitance. In this case, optimal voltage coupling takes place if the lead length is λ/2 (λ, wave length). An abandoned lead acts as an open circuit and has, thus, its optimum at λ/4. Determination of λ is rather difficult as the medium between the PG case and lead tip contains blood, muscles, lung tissue, and bones. For 64 MHz and blood, λ is 35 cm, for bone it is 90 cm. 11 As the distance between the lead tip and the case is decisive, this is about 20 cm, the connected lead is probably nearer to optimum coupling than the abandoned lead at that frequency. Luechinger 12 found that ‘the findings were similar when the leads were isolated or connected to pulse generators’. If one considers why leads are abandoned (rupture, insulation defect, and too high threshold), it is obvious that there is no risk of threshold elevation or other adverse effects at all. The listed adverse effects in the paper by Strach et al. are briefly commented on in the following. Two aspects evolve out of the critical analysis of the listed risks. First, if there is a realistic risk that pacemakers could be damaged or depleted by MRI as suggested by Strach et al.1 (also by other authors 8 ), MRI examination of pacemaker patients would be unethical because no safety measure could prevent it. Secondly, to program off the magnet mode and use the noise mode instead would increase safety. In the case of MRI incompatibility (inhibition or ventricular tachycardia), the supervising cardiologist could use the ‘red button’ to stop scanning and thereby the incompatibility. Further examination should then be carried out in the ECG-triggered mode as we proposed earlier. 9 Movement of pacemaker device or leads . Lead tips are never moved magnetically as they do not possess ferromagnetic materials. 9 , 13 The maximum force, according to Luechinger, is smaller than gravitation and are, thus, safe with respect to force effects. 12 Radiofrequency-induced heating of lead tip . This is a controversial topic as experimental results confirm heating of the metallic tip, whereas in vivo findings are scarce in demonstrating an intolerable increase in pacing threshold or histological damage. 13 Heating of the tip is not the deciding factor but heating within the closest layer of excitable tissue. Perforation of the myocardial wall can be excluded because the rapid declining temperature profile 12 and the power density never reach that of electrosurgery (up to 200 W). Pacemaker dysfunction such as induction of ventricular fibrillation. This risk is, indeed, the highest possible risk as it needs a considerable period of time to get the patient out of the magnetostatic field to defibrillate. This problem can possibly arise in patients with brady-tachycardia syndrome 9 and is not addressed in this paper. Asynchronous pacing together with increased magnetic rate can be dangerous as the fibrillation threshold decreases with increasing heart rate. 14 Damage . This has never occurred. Rapid pacing . This is no longer to be expected with modern PGs. Reed switch malfunction with or without asynchronous pacing . An irreversible malfunction has never been observed. Reed switches can open in strong fields 12 but this effect should not be disadvantageous. Inhibition of pacing output . This possibility was recently experienced in a 3 T machine 15 (see further remarks below). Change or loss of programmed data caused by an electrical reset . This effect does not constitute a real impairment during MRI and can be reversed by post-MRI programming. Depletion of the pacemaker battery voltage . This has never occurred. Change in pacing capture threshold . This is the same effect as discussed under RF heating. High abandoned lead risk . This is not a risk at all. Back-up mode creating cardiac insufficiency in children due to reduction in rate or output . Although young children may have a considerably higher rate than the 70 min −1 in the back-up mode, this should be tolerable during the span of an MRI examination. If the threshold of a child is higher than the output amplitude in the back-up mode, MRI examination is contraindicated. But is that a realistic situation or only of academic interest? Conflict of interest: none declared.