Dynamic left ventricular assist device inflow obstruction.

A 72-year-old man with a history of rheumatic heart disease , prior bioprosthetic mitral and aortic valve replacement , severe nonischemic dilated cardiomyopathy (left ventricular [LV] ejection fraction 5%), and a history of ven-tricular tachycardia (VT) underwent implantation of a continuous flow LV assist device (LVAD; HeartMate II; Thoratec Corporation) as destination therapy for severe heart failure. The early postoperative period was unremarkable and he had substantial functional improvement with minimal cardiovas-cular limiting symptoms after postoperative recovery. After ≈3 years on LVAD support, he experienced a series of appropriate implantable cardioverter-defibrillator shocks for sustained VT refractory to antiarrhythmic therapy. He underwent VT ablation targeting the basal LV inferoseptum, but he continued to have recurrent VT. Because VT seemed to originate from a deep septal substrate, he underwent an alcohol septal ablation that was also unsuccessful. A week later, VT was terminated during a second endocardial ablation and was noninducible thereafter. He has had no further sustained VT or implantable cardioverter-defibrillator shocks for 10 months of follow-up. Approximately 4 years after the LVAD implantation, and 6 months after the last VT ablation, he presented with new symptoms of intermittent dizziness, near syncope, and exer-tional dyspnea. Transesophageal echocardiography demonstrated a relatively improved LV ejection fraction of 25% with septal asymmetry and dynamic obstruction of the LVAD inflow with increased inflow velocities during each native heart systole. He was admitted and underwent repeat contrast echocardiograms to evaluate the obstruction further in varying positions and LVAD RPM settings (Figure; Movies 1–3 in the Data Supplement). A peak systolic velocity of 3.49 m/s was obtained with the patient upright at 9400 rpm. There was no clinical or laboratory evidence of pump thrombosis, including no changes in VAD parameters of power or pulsa-tility. The plasma lactate dehydrogenase concentration was 325 U/L (normal range 105–333) and the free hemoglobin level was 4.5 mg/dL (normal range 14–17). The material obstructing the LVAD inflow did not have thrombus-like quality on echocar-diographic imaging. Surgical interventions, including repositioning the inflow cannula with or without pump exchange and apical LV myo-mectomy, were considered but felt to carry excessively high risk. Because the LVAD inflow obstruction occurred mainly during native heart systole, we hypothesized that the physiology was similar to dynamic LV outflow obstruction as observed in hypertrophic obstructive cardiomyopathy. Moderate hydra-tion and adjustment of the LVAD rotor speed (RPM decrease from 9400 to 8800) to increase LV volumes and reduce inflow obstruction …