Karan Topiwala1, Haitham Hussein2, Kamran Masood2, Andrew Zhang3, Bhavani Kashyap4, Jason Bartos2, Huseyin Tore2, Eva Mistry5, Bentho Oladi2, Bharathi Jagadeesan2, Mustapha Ezzeddine6, Tapan Mehta7. 1. University of Minnesota, 516 Delaware St SE., 12-100 Phillips Wangensteen Building, Minneapolis, MN 55455, USA. Electronic address: topiwala.karan@gmail.com. 2. University of Minnesota, 516 Delaware St SE., 12-100 Phillips Wangensteen Building, Minneapolis, MN 55455, USA. 3. Cleveland Clinic, OH, USA. 4. Neuroscience Research, HealthPartners Institute, MN, USA. 5. Vanderbilt University, TN, USA. 6. Wake Forest Baptist Medical Center, NC, USA. 7. Ayer Neuroscience Institute, Hartford Hospital, CT, USA.
Abstract
OBJECTIVES: Recent case-reports have described an atypical cerebral microbleed (CMB) topography after extracorporeal membrane oxygenation (ECMO). The objective of this study was to examine the prevalence, radiographic patterns, and clinical correlates of possibly-ECMO-related (PER) CMB. MATERIALS AND METHODS: We performed a retrospective study of 307 consecutive patients receiving ECMO support at our tertiary-care University Hospital (2013-2018). PER CMB were defined as CMB present in corpus-callosum and/or middle cerebellar peduncle with/without involvement of other lobar/deep structures. Leukoaraiosis was quantified using the Wahlund age-related white matter changes scale. Patient characteristics were compared between cohorts with and without PER CMB. RESULTS: Forty patients (median age 60 years; 33% vv-ECMO and 67% va-ECMO) received at-least one MRI-brain within 3 months of ECMO support. CMB were present in 77.5% (n = 31) patients with 39% (n = 12), 17% (n = 5), and 44% (n = 14) having low (< 10 CMB), moderate (10-30 CMB), and high (> 30 CMB) burden respectively. Among CMB-positive patients, 71% (n = 22) had PER CMB, with 91% of such cases demonstrating involvement of splenium. Leukoaraiosis did not corelate to PER CMB presence (p = 0.267) or burden (ρ = 0.09). Patients with PER CMB had higher rates of ischemic stroke (50 vs. 33%), intracranial hemorrhage (41 vs. 17%), and all-cause mortality (27 vs. 17%); with survivors demonstrating no differences in their discharge disposition or modified Rankin Score. CONCLUSIONS: Post-ECMO cerebral microbleeds have a distinct distribution pattern that commonly involves the splenium of corpus-callosum. Their etiopathogenesis may be independent of microvascular lipohyalinosis. This requires further study in a larger sample-size.
OBJECTIVES: Recent case-reports have described an atypical cerebral microbleed (CMB) topography after extracorporeal membrane oxygenation (ECMO). The objective of this study was to examine the prevalence, radiographic patterns, and clinical correlates of possibly-ECMO-related (PER) CMB. MATERIALS AND METHODS: We performed a retrospective study of 307 consecutive patients receiving ECMO support at our tertiary-care University Hospital (2013-2018). PER CMB were defined as CMB present in corpus-callosum and/or middle cerebellar peduncle with/without involvement of other lobar/deep structures. Leukoaraiosis was quantified using the Wahlund age-related white matter changes scale. Patient characteristics were compared between cohorts with and without PER CMB. RESULTS: Forty patients (median age 60 years; 33% vv-ECMO and 67% va-ECMO) received at-least one MRI-brain within 3 months of ECMO support. CMB were present in 77.5% (n = 31) patients with 39% (n = 12), 17% (n = 5), and 44% (n = 14) having low (< 10 CMB), moderate (10-30 CMB), and high (> 30 CMB) burden respectively. Among CMB-positive patients, 71% (n = 22) had PER CMB, with 91% of such cases demonstrating involvement of splenium. Leukoaraiosis did not corelate to PER CMB presence (p = 0.267) or burden (ρ = 0.09). Patients with PER CMB had higher rates of ischemic stroke (50 vs. 33%), intracranial hemorrhage (41 vs. 17%), and all-cause mortality (27 vs. 17%); with survivors demonstrating no differences in their discharge disposition or modified Rankin Score. CONCLUSIONS: Post-ECMO cerebral microbleeds have a distinct distribution pattern that commonly involves the splenium of corpus-callosum. Their etiopathogenesis may be independent of microvascular lipohyalinosis. This requires further study in a larger sample-size.