Prognostic Implications of Delirium After Left Ventricular Assist Device (LVAD) Implantation: A Retrospective Study

Abstract: Background/Significance: Heart failure is a risk factor for delirium with a prevalence of up to 35%. While historically thought to be a benign disorder, delirium is now recognized as a condition prognostic for mortality and morbidity. This study investigates the impact of post-operative delirium (PODe) on measures of mortality and hospital morbidity after LVAD implantation.

Methods: We reviewed all adult patients receiving LVAD as bridge to transplant (BTT), Bridge To Decision (BTD), or destination therapy (DT) between June 2017 and December 2020 in a large health system in the Mid-Atlantic. We assessed risk factors for PODe; 1-month, 6-months, and in-hospital mortality; re-intubation rate; Length of Stay (LoS) in hospital and in ICU; discharge disposition and readmission rates.

Results: In total, 361 patients (26.7% women, 75.8% African American) received durable LVAD implants for BTT (24.4%), BTD (23.6%), and DT (51.9%). 94 patients (26.1%) developed PODe. Pre-implantation characteristics of gender, race, past medical or psychiatric conditions, INTERMACS score, and laboratory values did not differ between the two groups with and without PODe; age (55.88 vs 59.00; p= 0.031) was associated with development of PODe. Pre-operative use of inotropes and of vasopressors differed significantly between the two groups.

PODe was associated with higher 1-month (p=0.007), 6-months (p=0.004), in-hospital mortality (p< 0.001), and unplanned re-intubations (p< 0.001). Patients with delirium who survived their index admission were less likely to be discharged home (42% vs 60%, p = 0.03) and more likely to be discharged to a rehabilitation facility (43.6% vs 33.1%; p< 0.001). There was a clinically significant association of PODe with a longer LoS in hospital by 4.61 days, a longer LoS in ICU by 2.5 days, and earlier re-admission by 5.4 days. Delirium was not associated with a higher number of re-admissions.

A subgroup analysis shows that the treatment of PODe with either physical or chemical restraints in the post-LVAD population is associated with adverse hospitalization outcomes.

Discussion: In this single-center retrospective study, the prevalence of PODe was comparable to the prevalence previously reported in the literature. PODe was associated with higher mortality, adverse hospitalization outcomes, and unfavorable discharge disposition. This study highlights the need to screen for and aggressively prevent delirium in the LVAD surgical population, with particular emphasis on non-pharmacological strategies. Delirium also serves as a prognostic marker in this population and can inform clinical decision-making and discussions with patients and families regarding prognosis and recovery expectations.

Conclusion: Delirium is under-reported and under-detected in clinical settings. In the post-LVAD population, both hyperactive and hypoactive delirium are associated with higher mortality, and adverse outcomes. Future prospective studies are needed to identify modifiable delirium risk factors and establish the efficacy of specific delirium prevention strategies in this population.