Left ventricular assist device (LVAD) thrombosis is associated with high morbidity and mortality because of device malfunction, embolic events, and hemolysis. There remains uncertainty as to whether immediate device exchange versus an initial trial of anticoagulant and antiplatelet therapy is most appropriate in hemodynamically stable patients. We conducted a retrospective analysis of all LVAD implantations at a single center between January 2009 and June 2013 with follow-up through December 2013. Suspected LVAD thrombosis occurred in 20% of patients (N = 25) over a median follow-up of 275 days. Medical therapy led to resolution of hemolysis, and discharge to home, in 15 of 25 (60%) cases; however, this strategy was associated with intracranial hemorrhage in 4 patients and readmission with recurrent thrombosis in 10 patients. The 30 day, 6 month, and 1 year freedom from suspected LVAD thrombosis was 96.5, 85.9, and 80.3% in HeartMate II devices and 100, 92.9, and 87.1% in HeartWare ventricular assist devices, respectively (p = 0.11). Although medical treatment with intravenous heparin, antiplatelet agents, antithrombotic agents, or thrombolytic therapy can lead to initial resolution of hemolysis, the risks of recurrence after transition to warfarin and adverse events are high.