Cytokine adsorption in a patient with severe coronavirus disease 2019 related acute respiratory distress syndrome requiring extracorporeal membrane oxygenation therapy: A case report

Abstract

In serious novel coronavirus disease 2019 (COVID-19) with very severe acute respiratory distress syndrome (ARDS), even mechanical ventilation cannot provide satisfactory blood oxygenation and veno-venous extracorporeal membrane oxygenation (V-V ECMO) is necessary.1 Mortality in these patients is high, though reliable data are scarce. Current evidence describes mortality rates between about 30% and more than 90%.2-5 Elevated interleukin-6 (IL-6) levels are common in these serious cases and associated with poor outcome.6 Emerging experience and evidence indicate a potential benefit of managing elevated IL-6 levels in patients suffering from COVID-19. One approach to remove IL-6 from the circulating blood is extracorporeal adsorption within a specified cartridge (CytoSorb; CytoSorbents Europe, Berlin, Germany), which can be incorporated in extracorporeal circuits, like continuous veno-venous hemodialysis or ECMO. While there are reports on the control of the cytokine response by receptor blockage, there is no evidence on the use of cytokine adsorption in patients with COVID-19-related ARDS requiring ECMO therapy, so far.7 Therefore, we here report a case of a patient with severe COVID-19-related ARDS, where treatment with V-V ECMO and cytokine adsorption resulted in inflammation control and hemodynamic stabilization within 72 hours. A 59-year-old woman developed nausea, dizziness, and abdominal pain. Except for a history of breast cancer with ongoing hormone therapy and glaucoma, she did not suffer from any other relevant preexisting conditions. She presented herself to a family doctor who diagnosed a urinary tract infection after performing urine chemistry and prescribed nitrofurantoin and metamizole. Despite antibiotic therapy, the symptoms did not improve properly. Seven days later, she developed increasing dyspnea, however, without coughing nor fever. On day 8, she was finally admitted to a hospital with severe respiratory insufficiency requiring immediate noninvasive ventilation. By then, her body temperature was 38.8°C. Contrast-enhanced computed tomography images depicted multiple subsegmental pulmonary arterial emboli on both sides and signs of atypical pneumonia. Abnormalities in blood chemistry were elevated D-dimers, lymphopenia, and neutrophilia as well as increased levels of lactate dehydrogenase, C-reactive protein (CRP), and elevated cardiac biomarkers (troponin I hs, NT-proBNP); IL-6 was also altered. Renal and liver function parameters were only mildly elevated (Table 1). A polymerase chain reaction test for respiratory viruses finally revealed an infection with severe acute respiratory syndrome coronavirus 2. Hydroxychloroquine was administered for estimated antiviral effects according to an internal treatment standard for COVID-19 at that time and empiric antibiotic therapy (piperacillin/tazobactam and clarithromycin) was started for suspected bacterial superinfection. Additionally, she received anticoagulation therapy with unfractionated heparin monitored by partial thromboplastin time (PTT), aiming at a PTT of 60-80 seconds. However, the patient developed progressive hypoxemia (pO2 62 mm Hg under FiO2 100% on noninvasive ventilation) requiring invasive mechanical ventilation later that day. Prone positioning was started immediately. Echocardiography revealed mild to medium grade reduced ejection fraction. Norepinephrine (0.2 µg/kg/min) was started, and dobutamine was applied according to Pulse Contour Cardiac Output-measurements. Within the next days, CRP further increased. D-dimers and IL-6 dropped at first, but then showed a remarkable increment (Figure 1). Hypoxic respiratory failure worsened despite prone positioning and soon highly invasive ventilation was necessary (FiO2 100%, positive end-expiratory pressure [PEEP] 17 cm H2O, pmax 32 cm H2O). Due to the continuous deterioration, the patient was finally transferred to our center for initiation of V-V ECMO support 5 days after initial hospital admission. Additionally, antiviral therapy was escalated by adding ritonavir and lopinavir (following an internal treatment standard at that time). Laboratory findings immediately before implantation of ECMO showed elevated IL-6 levels (540 pg/mL) and we decided to integrate a CytoSorb cartridge into the ECMO circuit (SCPC Sorin, Munich, Germany) from the get-go. Due to a thrombus in the right jugular vein, a trans-jugular approach using a double-lumen cannula was not feasible and bi-femoral cannulation was performed. The CytoSorb adsorber was used according to the manufacturer's instructions and connected in a recirculating bypass starting behind the oxygenator and going back into the system at a pre-pump-luer-lock connection. For this setup, in continuous flow measurements, we detected blood flow rates through the adsorber of 350-450 mL/min, and thus, just within the range advised by the manufacturer (150-700 mL/min). Cytokine adsorption was continued for a total of 72 hours according to internal standards. Potential negative effects of the CytoSorb adsorber may be the adsorption of beneficial molecules like, for example, antibiotics. Based on available evidence, no relevant adsorption of the medication of our patient had to be suspected.8 After initiation of ECMO support and cytokine adsorption, we soon observed clinical stabilization of our patient. ECMO support with blood flow rates around 4 L/min allowed us to perform lung-protective low-tidal-volume ventilation as suggested by guidelines for ARDS treatment; to achieve PaO2 values between 55 and 65 mm Hg we required PEEP of 15 cm H2O and a fraction of inspired oxygen of 40%-60%. The need for vasopressors decreased significantly—dobutamine could be stopped shortly after initiation of ECMO and noradrenaline could be reduced step-wise and was no longer required after 4 days; lactate levels were never elevated during ECMO treatment. Also, CRP and IL-6 dropped within a few hours after the start of ECMO and cytokine adsorption while liver and renal function remained unaffected (Figure 1A-C, Table 1). From the beginning on, we observed abnormal clotting in the ECMO circuit, even under PTT-monitored anticoagulation with heparin or argatroban within in the target range (PTT 60 seconds) and various parts of the system and the cytokine adsorber needed to be replaced during the treatment. We observed a marked increase of D-dimers, while platelets decreased by more than 50% (Figure 2A,B, Table 1). Therefore, we changed from heparin to argatroban for suspected heparin-induced thrombocytopenia. However, a test for heparin-antibodies was negative. During the further course of treatment the patient developed a septic shock with multi-organ failure, most likely due to bacterial superinfection of the lung (procalcitonin levels ranged between 0.41 and 0.76 ng/mL (reference range: <0.05 ng/mL) during the first days after ECMO implantation and jumped to 18.4 ng/mL as a maximum on day 6 after ECMO implantation to remain markedly increased until death). According to the presumed will of the patient, the therapy was, therefore, terminated 12 days after initiation of ECMO and 17 days after initial hospital admission. We here report the case of a patient with severe ARDS due to COVID-19 treated with V-V ECMO and cytokine adsorption. This case suggests that cytokine adsorption may help initial stabilization of patients with severe COVID-19 disease requiring V-V ECMO support. Experience in COVID-19 patients requiring V-V ECMO support is limited and evidence on the outcome of these patients remains scarce. According to first reports, mortality in these serious cases is rather high, though available data are contradicting. While a pooled analysis of early reports revealed a mortality as high as 94.1%, others report lower rates around 30%-50%, which is consistent with preliminary results from the EuroELSO registry reporting on more than 1300 ECMO treatments, so far.3-5, 9 It is crucial to find additional therapeutic strategies to improve the outcome in these critically ill patients.10 According to recent evidence, elevated IL-6, reduced peripheral lymphocytes, and abnormal coagulation parameters were associated with poor outcome.6, 11 In our patient, we observed all these negative predictors. Therefore, we decided to integrate a CytoSorb cartridge in the V-V ECMO circuit as cytokine adsorption has been described as an approach to improve clinical outcomes in critically ill patients.12, 13 Previously, decreased vasopressor need and rapid hemodynamic stabilization after onset of cytokine adsorption was described for septic patients and also in COVID-19.14, 15 Consistent with these reports, over the short term, we observed hemodynamic stabilization during cytokine removal; the vasopressor need and levels of inflammatory parameters decreased significantly. However, based on this single case observation, we cannot say to what extent these effects were caused by cytokine adsorption or were merely a sign of general clinical stabilization of the patient, even if this occurred independently. While cytokine adsorption was effective and seemed to support stabilization of the patient, we observed severe clotting in the ECMO circuit even under full anticoagulation therapy as a serious therapeutic challenge. Hypercoagulability in COVID-19 has been described previously, and therefore, this complication is most probably not a consequence of cytokine adsorption or ECMO, but rather explained by the underlying disease.16, 17 The combination of V-V ECMO and cytokine adsorption may be feasible. Using the adsorber in a severe COVID-19 patient on V-V ECMO, we observed a considerable reduction of IL-6 and a hemodynamic stabilization. However, hypercoagulability leading to clot formation in the ECMO circuit required special caution. This experience from a single case treatment should be examined in more detail in larger cohorts. A Supady and D Duerschmied have received speakers’ honoraria from CytoSorbents Europe. A Supady received a research grant from CytoSorbents Europe. C Benk is shareholder and part-time employee of Resuscitec GmbH. The other authors declared no conflicts of interests. All authors agreed to publish this manuscript and have read and approved the final version.

Topics & Concepts