Mechanical Circulatory Assist Devices Stroke Subtype Classification: A Novel Stroke Classification System in Patients with Ventricular Assist Devices.

Background
Stroke is a frequent complication of left ventricular assist devices (LVADs) yet remains without a unified etiologic framework to guide evaluation and management. To address this, we developed a novel mechanical circulatory assist device stroke classification system with four etiologic categories: Management-related (M)-events driven by anticoagulation/antiplatelet strategy or perioperative management; Clinically related (C)-events attributable to conventional vascular risk factors; Acquired (A)-events resulting from LVAD-associated complications such as infection or coagulopathy; and Device-related (D)-events linked to pump- or graft-specific pathology or parameter abnormalities. We aimed to describe stroke subtype patterns and evaluate interrater reliability of this classification in LVAD patients.
Methods
We retrospectively reviewed 192 LVAD recipients at the University of Chicago Medical Center (2010-2020) and identified 32 neuroimaging-confirmed cerebrovascular events. Clinical characteristics, stroke subtype, contributing mechanisms, and outcomes were recorded. Two independent vascular neurologists classified each event using the MCAD system, and interrater reliability was assessed using intraclass correlation coefficients (ICC).
Results
Among the 32 cerebrovascular events, 59.4% were ischemic and 40.6% hemorrhagic (25% intraparenchymal, 12.5% subarachnoid). Strokes occurred a mean of 27 months after LVAD implantation. Strokes were categorized as Management-related (62.6%), Acquired (15.6%), Device-related (15.6%), and Clinically related (6.2%). Interrater agreement was excellent (ICC = 0.974).
Conclusion
This novel MCAD stroke classification demonstrated excellent interrater reliability in a single-center LVAD retrospective cohort and provides a structured framework for categorizing etiologic contributors. Future prospective multicenter studies are needed to assess its clinical utility.