Thrombus on the inflow cannula of the HeartWare HVAD: an update.

BACKGROUND: The HeartWare HVAD (Medtronic, Minneapolis, MN) is a continuous-flow left ventricular assist device (LVAD) approved by the FDA in 2012 as a bridge to transplant in patients with end-stage left ventricular heart failure. The current inflow cannula has a smooth outer surface near the inflow edge and a sintered collar of titanium microspheres near the pump. A previous case series of HVAD patients bridged to transplant revealed thrombus on the outer surface of the inflow cannula in 8 of 8 patients, predominantly at the smooth-sintered interface, that was associated with a clinical stroke rate of 12.5%.

DESIGN: Cases of HVAD devices removed at the time of heart transplant were identified in the surgical pathology database. The gross and microscopic findings were reviewed along with clinical data.

RESULTS: A total of 22 patients with 24 HVAD implants diagnosed with dilated cardiomyopathy (13 patients), ischemic heart disease (4 patients), lymphocytic myocarditis (2 patients), hypertrophic cardiomyopathy (2 patients), and congenital valvular disease (1 patient) were included. Two patients received two HVADs to provide biventricular support. All patients received post-implantation anti-coagulation with an INR goal of 2 to 3. Gross pathologic examination revealed thrombi on the outer aspect of the HVAD inflow cannula in 23 of 24 devices (96%). The inflow cannula of the one device that did not develop thrombus was positioned such that the smooth-sintered interface was buried in the ventricular myocardium and not in contact with blood in the ventricular chamber. Complications during the period of device support included 9 thromboembolic events (41%) including 6 ischemic strokes (27%), 2 intracoronary thromboembolic events and 1 splenic infarct. Patients suffered strokes 4 to 174 days (mean 82) after HVAD placement and had thrombus on the inflow cannula ranging in size from 0.1-2.5 cm (axial), 0.4-4.5 cm (circumferential) and 0.1-0.5 cm (thickness). Histologic evaluation revealed bland, partially organized thrombi without evidence of infection. Other complications included driveline infections (9%), non-driveline related bacteremia (9%) and hemorrhage (5%).

CONCLUSIONS: We report here an extension of our original study to a total of 22 patients with 24 HVAD implants who were all successfully bridged to transplant. We validate the very high prevalence of thrombus around the HVAD inflow cannula, associated with a clinical thromboembolic event in over a third of the patients, the majority of which were strokes. The nidus for thrombus formation appears to be the smooth-sintered interface of the HVAD inflow cannula.