Nazarian et al. [1] evaluate in a prospective non-randomized trial the safety of a 1.5 T MRI protocol for patients with implanted cardiac devices (ICD and PM). 438 participants (54 % with PM and 46 % with ICD implanted, respectively,after the 1998 and 2000) were enrolled and underwent 555MRI examinations. Patients with a recent implant(\6 weeks), those with abandoned or epicardiac leads, and PM-dependent patients with an ICD were excluded.According to the experimental protocol, an asynchronous pacing mode (VOO/DOO) was programmed in the pacemaker-dependent patients, while an inhibited pacing mode(VVI/DDI) was used for the other patients. The ICD function of non-PM-dependent patients was disabled. During the MRI examination, blood pressure, electrocardiography, pulse oximetry and symptoms were monitored by a nurse with experience in cardiac life support and device programming who had an immediate backup from an electrophysiologist.Device variables including sensing, impedance, capture threshold and battery voltage were evaluated before MRI examination, immediately after MRI and after 3–6 months.Variations exceeding 30, 40 and 50 %for, respectively, lead impedance, sensing and capture threshold were considered as significant changes in lead performance.Baseline and immediate follow-up interrogations were performed in all 438. Long-term follow-up device variables were available for 266 patients (61 %). 3 of 438 patients(0.7 %) experienced acute power-on-reset events. None of them had device dysfunction during long-term follow-up(3–6 months). Right ventricular sensing and atrial, right and left ventricular impedances were reduced immediately after MRI. At long-term follow-up in 61 % of the patients,decreased right ventricular (RV) sensing and lead impedance,increased RV capture threshold and decreased battery voltage were noted. The observed changes did not require device revision or reprogramming. The distributions of changes in device variables were within the 20 % at baseline for most participants. Thoracic MRI sequences had a greater effect on device variables and were more likely to result in artifacts (for instance, image distortion,signal voids or bright areas and poor fat suppression).