Potential harmful effects of magnetic resonance imaging in pacemaker patients should not be underestimated

We read with great interest the article by Irnich et al.,¹ in which the authors address the question whether future pacemakers, improved for magnetic resonance imaging (MRI) resistance, are needed. The authors report, for the first time, a systematic investigation on the causes of fatalities that occurred during MRI investigation and reported by Legal Medicine Departments in Germany. In this respect, the authors discuss several potential interactions between pacemakers and MRI, such as heating effects, reed-switch closure in strong magnetic fields, and stimulation and influence of sensing due to the gradient fields. The authors conclude that a small modification of the MRI unit and programmability of the magnet function of the pacemaker is all that is needed to make pacemakers compatible with MRI examinations.

In our opinion, the authors underestimate three main risks: inadvertent changes in pacing mode, induced voltage due to gradient fields, and the potential heating effects around the lead tip.

We support the opinion of the authors that the reported fatalities associated with MRI scanning may have occurred because of inappropriate pacing and not because of inhibition of pacing, as none of these patients was pacemaker-dependent. In our opinion, this may be due to the unpredictable pacemaker mode, as shown by the authors. Therefore, we think that a predictable magnet mode in high magnetic fields is absolutely necessary to avoid competitive rhythms (fast pacing or inhibition). An ‘MRI mode’ should have no sensing function to avoid inappropriate pacing because of the induced noise during MR scanning. It would be preferable if the pacemaker would automatically detect the strong magnetic field of the MRI device and convert to a pre-defined MRI mode.

In terms of the potential risks of cardiac stimulation by the gradient fields, the authors conclude that this does not pose a risk. However, this may not always be true because the input circuits of pacemakers commonly contain, in addition to a resistance of ∼10 kΩ, a capacitance and potentially other semiconductor parts with nonlinear behaviour. These may rectify and conduct the induced current to higher values, which may eventually capture the heart. This potential problem may be avoided by re-evaluation of the input circuits by the manufacturers.

The authors also conclude that heating at the lead tip is not a real problem in MRI scanning of pacemakers. We absolutely disagree with this comment, as it is based on an incomplete and simplified model. Heating due to radiofrequency pulses shows a power density fall to the fourth power of the distance (r₀/r)⁴ from the lead tip. This is correct for the radiofrequency-power deposition; however, the conclusion that the temperature will also drop in the same manner is not correct because heat conduction has to be taken into account, as previously described in different publications.^2,3 The example used by the authors describes a temperature decrease by a factor of 16, 1 mm away from the tip of the lead, which would only be correct in the case of a thermal isolator. In human tissue, which is a good thermal conductor, it is estimated that a decrease by a factor 2–4 would occur over a distance of 3 mm.⁴ In our opinion, an accurate prediction of the heat distribution around the lead tip is impossible because of the strong influence of perfusion, thermal conductivity, and thickness of scar tissue. Taking this into account, a temperature increase of 10°C or more may occur, which turns heating into a real problem in MRI scanning of pacemaker patients. Heating-induced effects on pacing parameters, e.g. capture threshold changes, were reported by Martin et al.⁵ in patients with implanted pacemakers undergoing clinical MRI.

The proposed MRI scanning scheme restricted to the ventricular refractory period by the authors will certainly be beneficial to solve some of the potential risks but carries the disadvantage of significantly prolonging the procedure. The increased scan duration will reduce the specific absorption rate of the sequence and therefore reduce the heating effects. However, scanning limited to the refractory period may not even be possible in special protocols, such as for cardiac scanning, or protocols using steady state, such as balanced steady-state free precession (SSFP) sequences, but also for sequences that need a high temporal resolution like perfusion and first pass angiograms. In our opinion, in all these cases and with the most commonly used sequences, a reliable MRI mode is absolutely needed.

References