In-Hospital Left Ventricular Assist Devices Deactivation and Death Experience: A Single-Institution Retrospective Analysis.

Abstract

Best practices for left ventricular assist devices (LVADs) deactivation at end-of-life (EOL) have yet to be elucidated. We conducted a single-institution retrospective review of patients who died following LVAD deactivation between January 2017 and March 2020. Data were obtained from institutional databases and electronic health record and were analyzed using descriptive statistics. Fifty-eight patients (70% male, 70% African American, median age 62 years) were categorized by implant strategy: bridge therapy (BT, N = 22, 38%) or destination therapy (DT, N = 36, 62%). Clinical events leading to deactivation were categorized either acute ( e.g. , stroke [ N = 31, 53%]), gradual decline ( N = 12, 21%), or complications during index hospitalization ( N = 15, 26%). Implant strategy was not associated with clinical trajectory leading to EOL ( p = 0.67), hospital unit of death ( p = 0.13), or use of mechanical ventilation ( p = 0.69) or renal replacement therapy ( p = 0.81) during terminal hospitalization. Overall time from admission to code status change was mean 27.0 days (SD 30.3 days). Compared with BT patients, DT experienced earlier do-not-resuscitate (DNR) orders ( p ≤ 0.01) and shorter survival post-deactivation ( p ≤ 0.01). Deactivations after gradual decline tended to occur outside ICUs, compared with acute events or index implant-related complications ( p = 0.04). Implant strategy was not associated with differences in EOL experience except regarding timing of DNR order and survival post-deactivation.