Background This analysis of the REMATCH Trial focuses on infection, which was an important source of morbidity and mortality. We use the information to suggest ways to decrease the incidence and effects of device-related infection. Methods Patients were randomized prospectively to receive left ventricular assist devices (LVADs) or optimal medical management (OMM) for end-stage heart failure. Infection variables included sepsis adjudicated as the cause of death; sepsis reported as a serious adverse event; percutaneous site or pocket infection; and pump housing, inflow- or outflow-tract infection. We compared the incidence and prevalence of events between groups and generated time-related descriptions. Results Survival with LVAD (n = 68 patients) was superior to OMM survival (n = 61 patients) with a 47% decrease in risk of death (p < 0.001), but the aggregate adverse event rate was greater for patients with LVADs (risk ratio, 2.29; 95% confidence interval, 1.85-2.84). Freedom from sepsis in patients with LVADs was 58% at 1 year and 48% at 2 years after implantation with superior survival in non-septic patients (60% vs 39% at 1 year and 38% vs 8% at 2 years in non-septic vs septic patients with LVADs, p < 0.06). Percutaneous site or pocket infection did not affect survival (p = 0.86). The hazard for onset of sepsis peaked within the first 3 weeks after implantation. Conclusions Survival is improved with permanent LVAD implantation compared with OMM therapy. However, infection causes substantial morbidity and mortality. Decreasing infections will increase survival and decrease morbidity in permanent LVAD recipients and will improve the risk-benefit ratio for permanent LVAD therapy. Copyright © 2004 by the International Society for Heart and Lung Transplantation.