Feasibility of pacing with digital twins of brainstem circuits to restore heart rate variability in heart failure

Background Small neural networks known as central pattern generators (CPG) are being increasingly recognized to re-establish physiological heart rate feedback response in pathologies such as heart failure. This approach restores the natural heart rate variability driven by neuronal autonomic tone that is lost in disease or with age. Our novel 2-lead neuronal-network pacemaker is developed to train and restore the physiologic pacing in adaption to breathing, arterial pO2, and blood pressure. In our first-generation pacemaker we were able to induce a paced respiratory sinus arrhythmia (RSA), mimicking a physiologic pattern. (Figure 1) We performed large-animal experiments to test the left ventricular hemodynamics in 2-chamber pacing with and without RSA, in healthy pigs and heart failure model after acute myocardial infarction. Methods A novel 2-chamber pacemaker with a diaphragm respiratory sensor was implanted in healthy pigs (n=5), while conventional single-chamber pacemaker was used in pigs with chronic anterior myocardial infarction (AMI) (n=4) and reduced ejection fraction (44±2,3%). Echocardiographic measurements were performed in both groups at baseline and during variable pacing mimicking the dynamic heart rate sensor at the AMI pigs, with/without preserved RSA at the healthy pigs. Cardiac output (CO) and stroke volume (SV) measurements were performed and compared between the groups. Results Constant pacing of the right ventricle without RSA with variable heart rate sequentially decreased the stroke volume in the AMI pigs from 45.1 (42.8-46.4) mL at the baseline to 42.9 (31.5-46.1) mL at 160bpm. (Figure 2) In the healthy pigs pacing without RSA decreased the stroke volume from 72.35 (69.66-74.42) mL to 56.75 (55.71-59.81) mL. By using our novel 2-lead neuronal-network pacemaker, the increasing RSA modulation during pacing sequentially increased the stroke volume (from 56.75 (55.71–59.81) mL to 63.04 (52.57–65.26) mL at 50% RSA and to 68.68 (68.23–73.2) mL at 100% RSA) at the healthy pigs. Conclusions Pacing with dynamic heart rate variability significantly decreased the stoke volume in AMI pigs with reduced left ventricular ejection fraction, presenting an unmet need to optimize hemodynamics in pacing patterns with dynamic heart rate sensor. In healthy pigs, pacing without RSA decreased the stroke volume, while adding the RSA modulation by our neuronal-network pacemaker improved the stroke volume.