Venovenous extracorporeal membrane oxygenation (VV ECMO) is used in cases of refractory respiratory failure, temporarily replacing function of the failing lungs. Use of VV ECMO is increasing along with a wider array of cannulas and devices available for use.1 Dual lumen cannulas have been developed as an alternative to the traditional use of two single lumen cannulas. Potential benefits include single access site and easier mobilization.2 The ProtekDuo right atrium (RA) to pulmonary artery (PA) dual lumen cannula was introduced in 2014 allowing percutaneous oxygenated right ventricular (RV) support (LivaNova, London, United Kingdom). However, in our experience, this cannula can be flow-limited by inadequate venous drainage requiring placement of an additional inflow cannula. We describe a series of four patients treated with ECMO for COVID-19–associated respiratory failure utilizing ProtekDuo cannula with an additional femoral inflow cannula complicated by ProtekDuo intracannula thrombus (Table 1). To our knowledge, this complication has not before been described in the medical literature. We discuss possible causes and need for further investigation.
Table 1. -
Description of Cannulation, ECMO Flows, and Anticoagulation Status Surrounding the Time of Thrombus Formation for Each Patient
| Patient |
Total ECMO Days |
ECMO Inflow |
ECMO Outflow |
Avg ECMO Flows (LPM) |
A/C Strategy |
PTT Goal (% at Goal) |
Avg PTT |
A/C Held (Total Days) |
Thrombus Formed (Day) |
| 1 |
76 |
25 Fr femoral multistage and proximal limb of 31 Fr ProtekDuo |
Distal limb of 31 Fr ProtekDuo |
3.98 |
Bivalirudin |
40–50 (49.5%) |
46.2 |
21 |
28 |
| 2 |
95 |
25 Fr femoral multistage and proximal limb of 31 Fr ProtekDuo |
Distal limb of 31 Fr ProtekDuo |
4.69 |
Bivalirudin |
50–60 (34%) |
56.8 |
23 |
29 |
| 3 |
66 |
25 Fr femoral multistage |
Proximal and distal limbs of 31 Fr ProtekDuo |
4.37 |
Bivalirudin |
55–65 (98%) |
62.8 |
4 |
50 |
| 4 |
69 |
25 Fr femoral multistage |
Proximal and distal limbs of 31 Fr ProtekDuo |
5.17*
|
Bivalirudin |
55–65 (92.8%) |
58.7†
|
0 |
3 |
ECMO Inflow and ECMO Outflow: ECMO cannulation strategy at time of thrombus formation. Avg ECMO flow: average total ECMO flows for 7 days* prior to thrombus formation. A/C strategy: type of systemic anticoagulation. PTT goal: target PTT at time of thrombus; in parentheses is the percentage of time on this platform that the patient was therapeutically anticoagulated. Avg PTT: average PTT for the 7 days† preceding thrombus formation. A/C held (total days): total number of days anticoagulation was held from initiation of ProtekDuo cannulation to thrombus formation. Thrombus formed (day): number of day that thrombus formed after initiation of ProtekDuo cannulation.
*For patient 4, avg ECMO flows is calculated from ProtekDuo placement to clot formation.
†For patient 4, avg PTT is calculated from ProtekDuo placement to clot formation.
ECMO, extracorporeal membrane oxygenation; LPM, liters per minute; PTT, partial thromboplastin time.
Patient Information
Patient 1
Forty-four-year-old male with no past medical history (PMH) presented to outside hospital (OSH) with COVID-19 pneumonia. Despite receiving baricitinib, dexamethasone, and remdesivir, his respiratory status declined. Failing invasive mechanical ventilation due to hypoxemia, he was cannulated by our mobile ECMO team with a 25 French (Fr) multistage right femoral venous inflow and 19 Fr single-stage right internal jugular (IJ) outflow. Cannulation was complicated by cardiac tamponade due to wire injury. He was transiently converted to venovenous arterial ECMO with 16 Fr left femoral arterial outflow cannula and underwent emergent pericardial window with subsequent conversion back to femoral-IJ VV ECMO the same day. He developed renal failure requiring continuous renal replacement therapy (CRRT) 10 days after cannulation. Worsening RV function prompted conversion to a right IJ 31 Fr ProtekDuo for additional RV support with right atrial (RA) proximal lumen inflow and PA distal lumen outflow 25 days after initial cannulation. To maintain adequate circuit flow while removing volume with CRRT, a 25 Fr multistage right femoral inflow was added 5 days later. Systemic anticoagulation was intermittently held following ProtekDuo placement due to oral bleeding and procedures, including tracheostomy. Twenty-eight days after ProtekDuo placement, thrombus was noted in the cannula inflow lumen. The inflow limb of the ProtekDuo was clamped while maintaining the patient on femoral inflow and PA lumen outflow. Clot was aspirated from the cannula and flow through this limb was reestablished. Following thrombus aspiration, the femoral inflow cannula was also removed. The patient’s renal function recovered but he showed no signs of lung recovery. He was ultimately deemed not to be a lung transplant candidate. At the patient’s request, ECMO support was terminally discontinued, and his kidneys were procured for organ donation after his passing.
Patient 2
Twenty-nine-year-old female with PMH Charcot-Marie-Tooth disease presented to OSH with altered mental status and hypoxemia due to COVID-19 requiring intubation on admission. Despite receiving sarilumab, dexamethasone, and remdesivir, she worsened and was cannulated for femoral-femoral VV ECMO with 25 Fr single-stage inflow and 25 Fr multistage outflow. She was extubated to high-flow nasal cannula and transferred to our facility for ECMO management. She was reintubated on arrival for respiratory distress. Despite treatment with proning, paralysis and ultra-low stretch mechanical ventilation, she exhibited minimal recovery of lung function. Twenty days after initial cannulation, she was converted to a 31 Fr right IJ ProtekDuo with RA inflow and PA outflow to assist with rehabilitation. An additional 25 Fr multistage femoral inflow cannula was added 3 days later given inability to maintain adequate flows. Systemic anticoagulation was held for 23 days after ProtekDuo placement due to ongoing airway bleeding and procedures including tracheostomy and chest tube placement. Severe hemolysis prompted conversion back to femoral-femoral VV ECMO with 25 Fr multistage inflow and 22 Fr single-stage outflow 29 days after ProtekDuo placement. At removal, thrombus was noted in the ProtekDuo inflow, thought to be the source of hemolysis. Acute RV dysfunction followed removal of ProtekDuo RV support. Placement of another oxygenated right ventricular assist device (oxy-RVAD) was planned; however, she was noted to have superior vena cava (SVC) syndrome due to SVC/RA thrombus that had formed around the prior ProtekDuo cannula. Catheter-directed lysis was unsuccessful. Four days after conversion to femoral-femoral ECMO, catheter-assisted clot retrieval was successfully performed. She was converted to 21 Fr right IJ single-stage outflow cannula positioned in the main PA with 25 Fr multistage right femoral inflow. Patient went on to successfully rehabilitate and underwent bilateral lung transplantation 95 days after initial cannulation. She was decannulated from ECMO intraoperatively and was discharged from the hospital 137 days after initial cannulation.
Patient 3
Thirty-year-old male with PMH anxiety presented to OSH with COVID-19 pneumonia. Despite receiving dexamethasone and baricitinib, he was intubated on hospital day 12. He acutely worsened after one session of plasmapheresis and was cannulated by our mobile ECMO team for right femoral–right IJ VV ECMO with 25 Fr multistage inflow and 19 Fr single-stage outflow on hospital day 16. Due to worsening RV function, he was converted to 31 Fr right IJ ProtekDuo (both lumens serving as outflow limbs) with 25 Fr multistage right femoral venous inflow on ECMO day 2. The RA lumen of the ProtekDuo was partially clamped to flow approximately 2.5 L/min through both limbs for total ECMO flows of 5 L/min. Following treatment with paralysis and proning, sedation was weaned and he was rehabilitated to the point of standing. On ECMO day 52, he was noted to have acute cessation of flow in the PA limb of ProtekDuo due to layering clot that had developed in the bend of the ECMO tubing (Figure 1). Notably, only total ECMO flow was being monitored at that time. At the time of PA limb flow cessation, total ECMO flows of 5 L/min were still achieved through the RA limb. An immediate echocardiogram ruled out intracardiac thrombus, and 19 Fr single-stage left IJ outflow cannula was placed with removal of ProtekDuo. No complications were noted and the patient transitioned from ECMO (after 74 days) to extracorporeal carbon dioxide removal (ECCO2R) and weaned from all support 81 days after initial cannulation.
;)
Figure 1.: Layering clot extending around bend in circuit tubing (yellow arrows) of distal/PA limb of ProtekDuo cannula taken at time of recognition. PA, pulmonary artery.
Patient 4
A 36-year-old female, who was 27 weeks pregnant, presented to OSH with COVID pneumonia and intubated 10 days after admission despite receiving baricitinib, remdesivir, sarilumab, and solumedrol. She transferred to our hospital and was cannulated for VV ECMO on arrival with a 25 Fr multistage right femoral inflow and 17 Fr single-stage right IJ outflow. Unfortunately, shortly after cannulation and fetal monitoring, she was noted to have intrauterine fetal demise and labor was induced. Labor was complicated by mild diffuse intravascular coagulopathy, which resolved after delivery. She was anticoagulated on bivalirudin 1 day later. Due to RV dysfunction, her cannulation was switched to a 25 Fr multistage right femoral inflow and right IJ 31 Fr ProtekDuo with both lumens functioning as outflow, 25 days after initial cannulation. Total flow was maintained at 5 L/min with 2.5 L/min flowing through each limb. Three days after cannula reconfiguration, small amount of fibrin buildup at the connection between circuit tubing and the ProtekDuo PA limb was noted, which was monitored closely. Systemic anticoagulation was maintained with bivalirudin throughout her ECMO course with exception of intermittent holding for procedures (chest tube placement, tracheostomy). After 15 days on this platform, rapidly increased volume of thrombus was noted on the minor curve of the circuit tubing leading to ProtekDuo PA limb (Figure 2). She was converted to left IJ 19 Fr single-stage outflow with existing right femoral 25 Fr inflow and ProtekDuo was removed. She was maintained on this configuration for 15 days before converting to left femoral 25 Fr multistage inflow and right IJ 21 Fr single-stage outflow positioned in main PA for RV support. She was rehabilitated on this platform, transitioned to ECCO2R 69 days after initial cannulation, and weaned from all support 4 days later.
Figure 2.: Extracorporeal membrane oxygenation (ECMO) circuit tubing traveling to distal limb of ProtekDuo, being used as a dual outflow with femoral inflow. Yellow arrow identifies adherent thrombus accumulating on wall of tubing in bend between shoulder and neck. Orange arrow indicating the direction of blood flow within both lengths of tubing traveling from ECMO oxygenator to internal jugular vein.
Discussion
Thrombus formation is a known risk of ECMO therapy. Most literature focuses on the formation of intravascular and oxygenator thrombus, rather than within circuit tubing and cannulas. This may in part be due to difficulty visualizing thrombus buildup within a functioning system, but incidence may also vary by cannulation strategy. We describe four cases of intracannula thrombus formation in patients supported with a dual lumen ProtekDuo cannula plus an additional femoral venous cannula for ECMO support.
ECMO disturbs the normal balance of hemostatic and anticoagulant factors through the introduction of complement, coagulation, and inflammatory cascade activation as well as consumption of coagulants, thrombocytopenia, and loss of von Willebrand (vWF) multimers.3 Furthermore, the foreign material surface of the ECMO circuit and blood flow alterations through and around access cannulas result in areas of turbulence and stasis, contributing to risk of thrombus as well as hemolysis.4 In addition, COVID-19 infection creates complex coagulation abnormalities leading to a hypercoagulable state. This further increases the risk of thrombus formation despite therapeutic anticoagulation.5 Bemtgen et al.6 observed higher rates of ECMO circuit thrombotic events in patients with COVID-19 compared with other causes of respiratory failure. These combined factors lead to greater risks of bleeding and thrombotic complications while on ECMO.
To minimize the risks of thrombus and secondary hemolysis, patients are typically maintained on systemic anticoagulation. Use of VV ECMO without use of systemic anticoagulation has been described, though is typically limited to shorter ECMO duration.3 Retrospective observation has found similar incidence of thrombus in patients maintained with and without systemic anticoagulation in a non-COVID population.3 Within the hypercoagulable COVID population, however, there are no reports of maintaining VV ECMO without systemic anticoagulation. While all four patients were maintained on systemic anticoagulation, periods of suspended anticoagulation may have contributed to accelerated thrombus formation. However, two of the four patients were therapeutically anticoagulated the majority of the time before thrombus formation (Table 1).
Thrombus formation within ECMO cannulas is a multifactorial process, which may be impacted by the cannula material and surface coating, biochemical conditions of the blood (balance of coagulation factors), and local flow conditions.7 Both flow stasis and high shear stress can lead to thrombus formation via different mechanisms. Areas of relatively stagnant flow contribute to fibrin production and adsorption to circuit tubing. This may occur along the minor curve of bends in circuit tubing.7 It has also been demonstrated that blood stasis occurs at the tip of multistage venous cannulas due to the majority of blood draining from the most proximal holes, furthest from the cannula tip.7 To date, there are no published computational fluid dynamics (CFD) models demonstrating areas of stagnancy (areas at risk for thrombus formation) or high velocities (areas at risk for hemolysis) in currently available double lumen RVADs with similar design to the ProtekDuo. However, CFD models of other dual lumen cannulas demonstrated areas of stagnancy in the curved areas of the cannula and at transitions from smaller to larger lumen diameters.8
While high flow velocities may seem protective against thrombus formation, high shear, low stasis environments may develop thrombus due to conformational change in vWF.9 Furthermore, circuit-induced hemolysis and circulating free hemoglobin may promote thrombosis within the ECMO circuit.10 Fibrinogen adsorption to blood-contacting materials creates a binding site for vWF, which may be potentiated by free hemoglobin. High shear forces unfold vWF, exposing binding sites and leading to platelet activation.2 Therefore, in the setting of high shear stress, the fibrinogen-bound vWF-free hemoglobin complex may become a catalyst for platelet aggregation and, thus, thrombus formation.
The ProtekDuo cannula consists of two distinct lumens, with the inner lumen entirely within the outer lumen, forming two concentric channels11 (Figure 3). We describe two different novel strategies of ProtekDuo use. The first is the traditional use of the dual lumen cannula with an additional multistage femoral inflow cannula in a (dl)VfVi-P configuration according to the ELSO Maastricht Treaty for ECLS nomenclature12 (Figure 4A). An oxy-RVAD configuration was chosen for patients given echocardiographic and biomarker evidence of significant RV dysfunction. An additional inflow cannula was added to improve drainage when adequate blood flow rates could not be maintained with ProtekDuo cannula alone. The two patients maintained with this (dl)VfVi-P configuration developed thrombus in the ProtekDuo RA-inflow limb (Figure 5, A and B). We noted that the size difference between the competing inflow paths (25 Fr femoral versus RA lumen of ProtekDuo) may contribute to thrombus formation in the (dl)VfVi-P configuration.
Figure 3.: Diagram of ProtekDuo cannula. Insert A demonstrates proximal inlet. Insert B demonstrating cannula cross-section at the level of proximal inflow holes showing concentric lumen (“cannula in a cannula”) design.
Figure 4.: Depiction of the ProtekDuo plus femoral cannula ECMO configurations utilized. A: ProtekDuo with additional femoral venous inflow in a (dl)VfVi-P configuration. B: Femoral venous inflow with both lumens of ProtekDuo serving as outflows in a Vf-(dl)ViP configuration. ECMO, extracorporeal membrane oxygenation.
Figure 5.: Locations of thrombus formation (yellow stars) within each patient’s circuit/cannula. A: Patient 1: thrombus formed in inflow lumen of the dual lumen cannula. Limb was clamped and tubing was disconnected from cannula to aspirate thrombus and reestablish flow. B: Patient 2: thrombus formed in inflow lumen/tubing of dual lumen cannula. C: Patient 3: thrombus formed in distal outflow lumen extending from bend in circuit tubing to the outlet in the main pulmonary artery. D: Patient 4: thrombus formed in the bend of circuit tubing flowing to distal outflow lumen.
Thus, the second strategy utilized a femoral inflow cannula with both lumens of the ProtekDuo cannula serving as outflows, in a Vf-(dl)ViP configuration12 (Figure 4B). Drawing on our experience of thrombotic complications in the (dl)VfVi-P configuration, the Vf-(dl)ViP configuration was adopted to provide RV assistance while still maintaining higher ECMO flows as high RVAD flows are associated with increased risk of pulmonary hemorrhage.13 This strategy allowed blood flow to be split between the right atrium and PA, ultimately providing partial RV support while limiting direct PA flows. In the two patients maintained with this Vf-(dl)ViP configuration, thrombus developed in the ProtekDuo PA-outflow limb (Figure 5, C and D).
Addition of a femoral inflow cannula has been described for patients supported with bicaval dual lumen cannulas without a reported increase in cannula thrombosis.14 A single case report has been published describing use of the Vf-(dl)ViP configuration without device complications.15 However, it is important to emphasize that the configurations we describe are off-label uses and to be cognizant of potential consequences of utilizing cannulas in these ways. A CFD model has yet to be published for either strategy of dual lumen cannula use, and due to our experience, we have stopped using the configurations described here. We hypothesize that in both configurations, the addition of the larger diameter femoral drainage cannula and creation of parallel blood flow paths contributes to altered flow dynamics within the ProtekDuo cannula, potentiating the risk of thrombus formation. Tables 2 and 3 detail the various configurations utilized among our COVID VV ECMO population. Notably, while additional femoral cannulas were required for other configurations, cannula-associated thrombus occurred more frequently in patients with ProtekDuo. In the setting of a hypercoagulable patient population, the risk for thrombus formation may be further exacerbated. How much COVID-associated coagulopathy contributed to thrombus formation, however, is unclear. Furthermore, it is unknown how the changes in PA and intrathoracic pressures affect fluid dynamics within the cannulas in an awake versus sedate patient. We suspect that increased pulmonary pressures in the setting of ARDS, coughing and participating in physical therapy, for example, may have detrimental effects on PA limb flows.
Table 2. -
Initial Cannulation Strategies Used for Patients With COVID ARDS Requiring VV ECMO
| Cannulation Strategy |
Number of Patients (Percentage) |
| Total patients |
108 |
| Initial: IJ-Femoral |
95 (88%) |
| Initial: Bicaval |
9 (8.3%) |
| Initial: Femoral-Femoral |
3 (2.8%) |
| Initial: ProtekDuo |
1 (0.9%) |
Initial cannulation strategies used included internal jugular-femoral cannulation, bicaval dual lumen cannula, femoral-femoral cannulation or ProtekDuo.
IJ, internal jugular; VV ECMO, venovenous extracorporeal membrane oxygenation.
Table 3. -
Details of ECMO Revisions
| Revision Required (Out of Total VV ECMO Patients, n = 108) |
Number of Patients (Percentage) |
| Total patients requiring revision |
50 (46.3%) |
| Required more than 1 revision |
14 (12.9%) |
| Type of Revision Required |
| Revised to bicaval dual lumen cannula |
35 (32.4%) |
| Required additional femoral inflow cannula |
5 (14.3%) |
| Developed cannula-associated thrombus |
0 |
| Revised to IJ-Femoral |
12 (11.1%) |
| Required additional femoral inflow cannula |
3 (25%) |
| Developed cannula-associated thrombus |
1 |
| Revised to ProtekDuo |
11 (10.2%) |
| Required additional femoral inflow cannula |
6 (60%) |
| Developed cannula-associated thrombus |
4 |
| Revised to other percutaneous oxy-RVAD (spectrum dual lumen cannula or femoral + PA single lumen cannulas) |
5 (4.6%) |
| Revised to V-VA or VV-A |
4 (3.7%) |
| Revised to femoral-femoral |
3 (2.8%) |
Revisions included changing cannulas due to infection, thrombus, hemolysis or changes in patient condition requiring a different cannulation strategy. Percentage of patients requiring revision is out of the total number of patients requiring VV ECMO (n = 108). Numbers in parentheses are the percentage of patients that required an additional femoral inflow cannula out of the total number of patients requiring that same cannulation strategy.
IJ, internal jugular; oxy-RVAD, oxygenated right ventricular assist device; VV ECMO, venovenous extracorporeal membrane oxygenation; VV-A, venovenous arteria.
Bold indicates the total number of patients in each group.
In summary, we describe a series of four patients with COVID ARDS supported with ProtekDuo dual lumen cannula associated with intracannula thrombus. To our knowledge, this is the first publication detailing thrombus formation within a dual lumen RA to PA ECMO cannula. In all four patients, thrombus developed in the setting of an additional venous cannula in two novel ECMO configurations. Further investigation is needed to understand the complex interplay of factors contributing to intracannula thrombus formation.
References
1. Extracorporeal Life Support Organization (ELSO). ELSO Registry. Ann Arbor, MI: ELSO, 2022.
https://www.elso.org/Registry/InternationalSummaryandReports/InternationalSummary.aspx. Accessed February 12, 2022.
2. Broman LM, Westlung CJ, Gilbers M, et al.: Pressure and flow properties of dual-lumen cannulae for extracorporeal membrane oxygenation. Perfusion 35: 736–744, 2020.
3. Olson S, Murphree C, Zonies D, et al.: Thrombosis and bleeding in extracorporeal membrane oxygenation (ECMO) without anticoagulation: A systematic review. ASAIO J 67: 290–296, 2021.
4. Ogawa S, Richardson JE, Sakai T, Ide M, Tanaka KA: High mortality associated with intracardiac and intrapulmonary thromboses after cardiopulmonary bypass. J Anesth 26: 9–19, 2012.
5. Helms J, Tacquard C, Severac F, et al.: High risk of thrombosis in patients with severe SARS-CoV-2 infection: A multicenter prospective cohort study. Intensive Care Med 46:1089–1098, 2020.
6. Bemtgen X, Zotzmann V, Benk C, et al.: Thrombotic circuit complications during venovenous extracorporeal membrane oxygenation in COVID-19. J Thromb Thrombolysis 51: 301–307, 2021.
7. Vatani A, Liao S, Burrell AJC, et al.: Improved drainage cannula design to reduce thrombosis in veno-arterial extracorporeal membrane oxygenation. ASAIO J 68: 205–213, 2022.
8. Condemi F, Wang D, Fragomeni G, et al.: Percutaneous double lumen cannula for right ventricle assist device system: A computational fluid dynamic study. Biocybern Biomed Eng 36: 482–490, 2016.
9. Casa LDC, Deaton DH, Ku DN: Role of high shear rate in thrombosis. J Vasc Surg 61: 1068–1080, 2015.
10. Valladolid C, Yee A, Cruz MA: Von Willebrand factor, free hemoglobin and thrombosis in ECMO. Front Med 5: 228, 2018.
12. Broman LM, Taccone FS, Lorusso R, et al.: The ELSO Maastricht treaty for ECLS nomenclature abbreviations for cannulation configuration in extracorporeal life support – A position paper of the Extracorporeal Life Support Organization. Crit Care 23: 36, 2019.
13. Abdelshafy M, Caliskan K, Guven G, et al.: Temporary right-ventricular assist devices: A systematic review. J Clin Med 11: 613, 2022.
14. Chimot L, Marqué S, Gros A, et al.: Avalon bicaval dual-lumen cannula for venovenous extracorporeal membrane oxygenation: Survey of cannula use in France. ASAIO J 59: 157–161, 2013.
15. Maybauer MO, Koerner MM, Mihu MR, Harper MD, El Banayosy A: The ProtekDuo as double lumen return cannula in V-VP ECMO configuration: A first-in-man method description. Ann Card Anaesth 25: 217–219, 2022.
