Cardiac computed tomography improves the identification of cardiomechanical complications among patients with suspected left ventricular assist device malfunction.

Journal Article (Journal Article)

BACKGROUND: Left ventricular assist devices (LVAD) are increasingly used for durable mechanical circulatory support in advanced heart failure. While LVAD therapy provides substantial improvement in mortality and quality of life, long-term therapy confers increased risk for device complications. We evaluated if cardiac computed tomography (CCT) improves the detection of cardiomechanical complications among patients with LVAD and suspected device malfunction. METHODS: In this study, we compared the diagnostic performance of CCT and transthoracic echocardiography (TTE) for the identification of cardiomechanical LVAD complications, including thrombus or neointimal hyperplasia, inflow cannula malposition with dynamic obstruction, fixed outflow obstruction, device infection, and severe aortic regurgitation. Complications were confirmed with surgical evaluation, pathologic assessment, or response to therapeutic intervention. RESULTS: Among 58 LVAD patients, who underwent CCT and TTE for suspected LVAD dysfunction, there were 49 confirmed cardiomechanical LVAD complications among 43 (74.1%) patients. The most common LVAD complication was thrombus or neointimal hyperplasia (65.3%), followed by dynamic obstruction (26.5%). Individually, CCT identified 29 of the 49 (59.2%) confirmed LVAD cardiomechanical complications, whereas TTE alone identified a complication in 11 cases (22.4%). However, diagnostic performance was greatest when the two modalities were used in combination, yielding a sensitivity of 67%, specificity of 93%, PPV of 97%, NPV of 47% and diagnostic accuracy of 73%. CONCLUSION: The novel and complementary use of CCT with TTE for the evaluation of suspected device malfunction improves the accurate identification of cardiomechanical LVAD complication compared to either modality alone.

Full Text

Duke Authors

Cited Authors

  • Patel, PA; Green, CL; Lokhnygina, Y; Christensen, J; Milano, CA; Rogers, JG; Patel, CB; Koweek, LM; Daubert, MA

Published Date

  • 2021

Published In

Volume / Issue

  • 15 / 3

Start / End Page

  • 260 - 267

PubMed ID

  • 32891544

Electronic International Standard Serial Number (EISSN)

  • 1876-861X

Digital Object Identifier (DOI)

  • 10.1016/j.jcct.2020.08.008


  • eng

Conference Location

  • United States