Abstract
BACKGROUND: Left ventricular assist device (LVAD) speed optimization and right heart failure post device implantation are major clinical challenges. Right heart catheterization (RHC)-guided speed titration studies are often performed to optimize LVAD settings, which are unknown and must be optimized for each patient. A virtual hemodynamic model (VHM) that can be tailored to each patient may provide useful guidance and reduce repeated studies. METHODS: We conducted a retrospective analysis on 16 patients implanted with HeartMate 3 (HM3) who underwent RHC speed titration study as an outpatient. A custom-designed VHM was built and customized for each patient based on RHC measurements. VHM predictions were obtained for multiple scenarios: (1) population-based pulmonary system parameters, (2) patient-specific systemic and pulmonary resistance and capacitance parameters, (3) clinical optimization-based patient-specific mean arterial pressure (MAP), and (4) several MAP targets ranging from 70 to 90 mm Hg. RESULTS: All patients who underwent RHC speed titration had a clinician-guided speed increase, with a median increase of 300 revolutions per minute (rpm). Using each patient's customized VHM, virtual speed optimization demonstrated congruence with clinician-guided optimization, with a median predicted speed increase of 321 rpm. After virtual optimization, there was a decrease in the pulmonary artery pressure for 13 patients (81.25%), indicating a predicted improvement in pulmonary parameters. CONCLUSIONS: For our cohort of 16 patients, there was an overall congruence between clinician-guided and patient-specific VHM-predicted optimal LVAD speeds. The magnitude of speed change varied depending on individual patient targets. This may provide individualized speed titration goals and lessen the need for repeat invasive studies.