In 2003, Medicare expanded coverage of ventricular assist devices as destination, or permanent, therapy for end-stage heart failure. Little is known about the long-term outcomes and costs associated with these devices.To examine the acute and long-term outcomes of Medicare beneficiaries receiving ventricular assist devices alone or after open-heart surgery.Analysis of inpatient claims from the Centers for Medicare & Medicaid Services for the period 2000 through 2006. Patients were Medicare fee-for-service beneficiaries who received a ventricular assist device between February 2000 and June 2006 alone as primary therapy (primary device group; n = 1476) or after cardiotomy in the previous 30 days (postcardiotomy group; n = 1467).Cumulative incidence of device replacement, device removal, heart transplantation, readmission, and death, accounting for censoring and competing risks. Patients were followed up for at least 6 months and factors independently associated with long-term survival were identified. Medicare payments were used to calculate total inpatient costs and costs per day outside the hospital.Overall 1-year survival was 51.6% (n = 669) in the primary device group and 30.8% (n = 424) in the postcardiotomy group. Among primary device patients, 815 (55.2%) were discharged alive with a device. Of those, 450 (55.6%) were readmitted within 6 months and 504 (73.2%) were alive at 1 year. Of the 493 (33.6%) postcardiotomy patients discharged alive with a device, 237 (48.3%) were readmitted within 6 months and 355 (76.6%) were alive at 1 year. Mean 1-year Medicare payments for inpatient care for patients in the 2000-2005 cohorts were $178,714 (SD, $142,549) in the primary device group and $111,769 (SD, $95,413) in the postcardiotomy group.Among Medicare beneficiaries receiving a ventricular assist device, early mortality, morbidity, and costs remain high. Improving patient selection and reducing perioperative mortality are critical for improving overall outcomes.