Reply to the letter of Vahlhaus

Dr Valhaus, praising and also criticizing our paper on ‘Do we need pacemakers resistant to magnetic resonance imaging?’¹ speculates that ‘cardiac arrest caused by sinus arrest in a setting of possible MRI inhibition of the pacemaker (PM) could also be the trigger leading to ventricular fibrillation (VF) on the basis of cardiac hypoxia’. This involves two assumptions: there must be a prolonged sinus node recovery time in a non-PM-dependent patient and simultaneously the PM must be inhibited by the MRI device. All stored electrograms in ICDs^2,3 demonstrate that MRI noise is continuous and, therefore, makes the devices switch to ‘interference mode’, which is asynchronous pacing in PMs and inhibition in ICDs. Thus, the deceased PM patients listed in our paper were certainly paced asynchronously either by interference mode or by magnet mode.

Comments on Vahlhaus's statements

1. The paper by Vahlhaus et al.⁴ states that: ‘While in two patients the sensing threshold increased immediately after MRI (7.2 vs. 8.5 mV; 8.0 vs. 11.2 mV), it decreased in two patients (4.0 vs. 3.2 mV; 7.0 vs. 4.0 mV). There was no significant difference in sensing thresholds immediately after MRI and at the three-month follow-up compared with baseline values.

   In one patient, atrial stimulation threshold increased by one step (atrial stimulation threshold at 2 V: from 0.2 ms impulse duration before MRI to 0.25 ms after MRI) immediately after MRI. The temporary increase in the atrial stimulation threshold during the first MRI examination was not observed again after the second or the third MRI’.

   If one considers the rather low accuracy and reproducibility of sensing and pacing threshold measurements by implanted PMs, it is rather speculative to assign the facts described to the heating of the electrode.

2. In the paper of Wollmann et al.,⁵Figure 3 represents changes in sensing and pacing thresholds. However, if the curves are analysed, one can but state that pre- and post-MRI data were nearly identical for all three trials and that a threshold increase from 0.8 to 1.6 ms in pulse duration is only marginal, if a modern electrode with a chronaxie of ≤0.2 ms can be assumed. In this case, the threshold increases from U_rheobase(1+0.2/0.8 ms) to U_rheobase (1+0.2/1.6 ms), which is an increase of not more than 11%. Threshold measurements far right of chronaxie are too inaccurate to make valid threshold comparisons.⁶ This means that all thresholds before and after MRI examination are practically identical.

3. The dysfunctions of the ICD that Anfinsen et al.² describe were (i) noise being detected as VF and (ii) prolonged charge time with the indication ‘end of life’. The threshold increased from 0.4 V, 0.5 ms at implantation to 1.6 V, 0.5 ms 2 weeks after MRI examination, which took place 8 days after implantation. No other threshold measurements were made. To interpret the threshold increase 22 days after implantation as deterioration due to heating is more than daring, as this change may have been related to post-implant evolution.

4. Temporary loss of capture for 12 h, as reported by Roguin et al.³ in one animal, must not conclusively be traced to thermal damage. Tissue heated to 50°C or more is irreversibly damaged by denaturization. Why should the threshold return to normal after 12 h if excitable tissue around the electrode is denaturized by heat? The authors interpret this as ‘some oedema occurring at the lead tip-tissue interface, which subsequently resolved.’ Are oedema and threshold elevations of screw-in leads only to be expected when RF fields are present?

5. Although Sommer et al.⁷in vitro and Luechinger et al.⁸in vivo found alarmingly high temperatures at the electrode tip, the investigation in six animals undergoing MRI examinations with high SAR values yielded results prompting the following comments from the authors: ‘heat-induced tissue damage at the histological level could not be seen with certainty in any of the acute or chronic experiments’ and ‘the observed changes in stimulation threshold were not likely to be clinically relevant’.⁸ Though the theoretical power density is only one parameter in the balance of heat production, it helps to explain why the electrode surface may be more than 50°C with only minor changes in threshold in tissue 1–2 mm away. To this may be added the remarks on heat balance in the reply to Luechinger et al. in point 3.

References