Influence of different electrode placement patterns on electromagnetic interference generated by Left Ventricular Assist Devices (LVADs) during electrocardiogram performance

Abstract

Abstract Background Heart failure (HF), with an incidence of approximately 3/1000 person-years, often progresses to advanced stages where symptoms persist despite treatment. In such cases, mechanical circulatory assist devices, particularly left ventricular assist devices (LVADs) in ambulatory patients, enhance survival and quality of life. These LVADs generate electromagnetic interference, resulting in noise artefacts in conventional electrocardiograms (ECGs). The conventional ECG should be configured with a low-pass filter of 150 Hz and a high-pass filter of 0.05 Hz, as recommended by the American Heart Association/American College of Cardiology/Heart Rhythm Society. LVADs, such as the HeartMate 3 (Abbott®), typically operate at a rotational speed of 5000 to 6000 rpm, corresponding to oscillatory frequencies of 83.3 to 100 Hz. Objective To assess the impact of various surface electrode placement strategies and modifications to the low-pass filter on the electromagnetic interference produced by LVADs. Methods We conducted an observational study involving seven patients with HeartMate 3 adjusted to various rotational speeds (refer to Table 1). Three distinct electrode placement schemes (AHA, Mason Likar, and modified Torso) were employed for the frontal plane leads. The horizontal plane leads remained unmodified. Each patient underwent a total of 6 ECGs, corresponding to each of the aforementioned schemes. ECG recordings were performed using a high-pass filter set at 0.05 Hz and a low-pass filter set at 150 Hz. Additionally, a set of ECGs was obtained by modifying the low-pass filter to 40 Hz for further analysis. Results Figure 1 displays ECGs obtained from a male with a HeartMate 3 operating at 4600 rpm or 76.6 Hz. This device employs an algorithm to generate an "artificial pulse" by cyclically accelerating and decelerating the pump speed by 2000 rpm every 2 seconds, resulting in rotational speeds of 2600 rpm and 6600 rpm (43.3 Hz and 110 Hz, respectively). No specific placement scheme demonstrated consistent noise reduction across all cases, likely due to inter-individual variability resulting from factors such as the final position of the LVAD the rpm in each case, and the thoracic impedance of each patient. However, the modification of the low-pass filter to 40 Hz yielded a noteworthy improvement in noise reduction across all cases. In our limited sample, this adjustment did not pose issues for ECG interpretation. Nevertheless, it is essential to acknowledge that excessively low filter settings may not only eliminate noise but also potentially discard clinically relevant high-frequency signals (e.g., pacemaker spikes, large amplitude Rs, QRS notches). Conclusions Individualizing the electrode placement scheme is crucial for minimizing noise generated by LVADs. The adjustment of the low-pass filter to 40 Hz could generally prove beneficial in reducing artifact generation, despite the clinical limitations it may entail.