RESUMO
INTRODUCTION: Patients who have ventricular assist devices (VADs) and experience bloodstream infection (BSI) have high mortality. We addressed 2 questions raised by the United Network for Organ Sharing (UNOS) priority policy for this problem: 1) Are organs wasted on this ultra-high-risk group? 2) Can device-related BSI be differentiated from transient BSI? METHODS: Patients with VADs who underwent heart transplantation from 1987 to 2001, who had BSI during VAD support, and who had positive cultures at VAD explant (device-related BSI, n = 10) were compared with those with negative cultures at explant (non-device-related BSI, n = 11). RESULTS: Patients with device-related BSI had an 80% (8/10) rate of persistent bacteremia; 30 days and 1 year after transplantation, mortality was 14% and 26%, respectively. Non-device-related BSI (n = 11) persisted in 18% (2/11); peri-operative and 1-year mortalities were 9% and 13%. Duration of VAD support predicted infection (132 vs 48 days, p < 0.001); hypo-albuminemia (2.9 +/- 0.5 mg/dl vs 3.3 +/- 0.8 mg/dl, p < 0.05), and a resistant organism predicted a device-related BSI. These patients had increased intubation requirements and had increased creatinine concentration during the first post-operative week, with no difference in liver function, blood loss, transfusions (packed red blood cells, fresh frozen plasma, or platelets), or hemodynamic stability vs patients with non-device BSI. Despite decreased immunosuppression, we found no difference in acute rejection events with device-related BSI. Re-infection with the pre-operative organism occurred in only 1 patient per group. CONCLUSIONS: These data suggest that urgent (Status 1A) cardiac transplantation is effective in stable patients with device-related BSI, and these data support the current UNOS policy. However, an extra-device source of BSI should be excluded by considering the isolated organism, the baseline nutritional status, and other risk factors.