Thrombosis inside the membrane oxygenator (MO) is a critical complication during venovenous extracorporeal membrane oxygenation (ECMO). The aim of this study was to prove if thrombotic clots manifest within the MO when D-dimer levels are elevated over a long-term period. Heparin-coated polymethylpentene MOs (n = 13) were exchanged due to high plasma D-dimer levels. Clot volume was calculated using multidetector computed tomography (MDCT). Coagulation parameters and MO function were analyzed before and after MO exchange. Before MO exchange, D-dimer levels increased significantly in each patient (11.5 [6.5-15.5] mg/L to 35.0 [34-35] mg/L, P ≤ 0.001). High levels of D-dimers were tolerated for 1 to 6 days. Additionally, fibrinogen concentration (n = 8) and platelet count decreased (n = 8). Within 48 h after exchange, D-dimer levels decreased significantly (n = 11, 12 [8-16] mg/L, P = 0.004). Fibrinogen concentration and platelet counts increased. Clots were found in all MOs in the inlet part of the device. Clot volume (16-106 cm(3) ) did not correlate with MO support time but increased significantly when high D-dimer levels were accepted for >2 days. An increase or high levels of D-dimers in absence of other explaining pathology during ECMO therapy reflected coagulation activity within the MO. Evidence of clots within the MO at high D-dimer levels and decrease after exchange underline the relevance of D-dimer testing during ECMO treatment. Besides, surveillance of MOs during ongoing ECMO therapy will help to predict clot formation, and to avoid system-induced coagulation disorders as well as critical situations.

Oxygenator thrombosis is a serious complication in extracorporeal membrane oxygenation (ECMO) and may necessitate a system exchange. Coagulation and fibrinolysis parameters, flow dynamics, and gas transfer performance are currently used to evaluate the degree of oxygenator thrombosis, but there is no technical approach for direct visualization and quantification of thrombotic deposits within the membrane oxygenator (MO). We used multidetector computed tomography (MDCT) with three-dimensional postprocessing to assess the incidence of oxygenator thrombosis, to quantify thrombus extent, and to localize clot distribution. Twenty heparin-coated MOs after successful weaning were analyzed. Mean ECMO support time was 7 ± 4 days, mean activated partial thromboplastin time (aPTT) during ECMO was 59 ± 20 seconds. Thrombotic deposits were detected in all MOs. The mean clot volume was 51.7 ± 22.3 cm. All thrombotic deposits were located in the venous, i.e., inlet part of the device, without apparent evidence of embolization in patients. There was no correlation between clot volume and ECMO support time or aPTT. Clot formation within the MO is a common finding in ECMO despite adequate systemic anticoagulation. The clinical significance of thrombus formation and its influence on gas exchange capacity and hemostatic complications have to be addressed in further studies.

Despite heparin coating and systemic anticoagulation, thrombotic clot formation is a serious complication in extracorporeal membrane oxygenation (ECMO). We describe our first results of visualization of thrombotic deposits in ECMO devices using advanced multidetector computed tomography (MDCT). A bioline-coated polymethylpentene membrane oxygenator (MO) after 8 days of ECMO treatment (device 1) and a factory-sealed MO serving as an internal quality control (device 2) were analyzed with three-dimensional (3D) visualization volume rendering technique (VRT) using a 0.6 mm3 voxel isotropic MDCT dataset. After the computed tomography (CT) scan, device 1 was anatomically dissected for direct visualization of potential deposits and further analyzed by scanning electron microscopy (SEM). The VRT 3D model based on the MDCT dataset of device 1 showed red-coded areas within the gas exchange surface of the device consistent with fibrous and cellular deposits. These deposits could be confirmed by anatomical dissection of the device and by SEM. Device 2 showed no signs of clot formation in MDCT using the same VRT settings. It was demonstrated that MDCT with VRT is able to detect thrombotic deposits in ECMO devices under ex vivo conditions. MDCT allows direct visualization of the actual thrombus load of a used ECMO device as well as the quantification of the thrombus volume and could, therefore, play a significant role in better understanding the oxygenator thrombosis in modern ECMO treatment.