Extracorporeal membrane oxygenation (ECMO) has established as a cornerstone therapy in severe acute respiratory distress syndrome and refractory hemodynamic failure. As circuit integrity is crucial for adequate organ support, component failure may necessitate a system exchange. In this retrospective study, incidence and etiology of system exchanges during applications of venovenous, venoarterial ECMO, and extracorporeal CO₂ removal were examined. Sixty-three (44.4%) of 142 patients were affected by one or more exchanges, totaling 105 replaced circuits. The predominant exchange reason was clotting (n = 20), followed by hemolysis (n = 19), systemic coagulation disorders (n = 13), reconfiguration (n = 13), impaired gas exchange (n = 10), mechanical complications (n = 8), bleeding (n = 6), failed weaning (n = 5), prophylactic exchange (n = 3), and undocumented/other (n = 8). Nineteen (18.1%) events were classified as acute and 70 (66.7%) events as elective exchanges. Patients with circuit exchanges more frequently underwent renal replacement therapy at ECMO initiation (49.2% vs. 29.1%; p = 0.023), had a longer ECMO treatment duration (18 vs. 7.5 days, p < 0.001), and lower hospital survival (29.5% vs. 57.1%; p = 0.002). Considering the high occurrence of coagulation complications, further optimization of coagulation management is deemed necessary.

Rationale: Prostaglandin E1 (alprostadil; PGE1), in addition to low-dose unfractionated heparin, increases the biocompatibility of extracorporeal systems and enhances the efficacy of artificial organs without increasing bleeding risk. Objectives: We investigated the safety and efficacy of PGE1 in adults receiving venovenous extracorporeal membrane oxygenation (ECMO). Methods: This study was a randomized, double-blind, placebo-controlled phase II pilot trial at two medical intensive care units at the Medical University of Vienna, Austria. Adults with venovenous ECMO were randomly assigned to receive an intravenous infusion of 5 ng/kg/min PGE1 or placebo (0.9% saline) in addition to standard anticoagulation with unfractionated heparin. Measurements and Main Results: The primary outcome was the rate of transfused packed red blood cells per ECMO day. Secondary outcomes were the incidence of and time to clinically overt bleeding and thromboembolic events. A post hoc subgroup analysis included only patients with coronavirus disease (COVID-19). Between September 2016 and April 2021, of 133 screened patients, 50 patients were randomized, of whom 48 received the assigned study medication (24 per group). The transfusion rate was similar between groups (0.41 vs. 0.39; P = 0.733). PGE1 was associated with fewer thromboembolic events (7 vs. 16; P = 0.020) and longer thromboembolism-free time (hazard ratio [HR], 0.302; P = 0.01), fewer clinically overt bleeding events (2 vs. 11; P = 0.017), and longer bleeding-free time (HR, 0.213; P = 0.047). In patients with COVID-19 (n = 25), the HRs for clinically overt bleeding and thromboembolism were 0.276 (95% confidence interval, 0.035–2.186) and 0.521 (95% confidence interval, 0.149–1.825), respectively. Conclusions: Add-on treatment with PGE1 was safe but did not meet the primary endpoint of reducing the rate of red blood cell transfusions in patients receiving venovenous ECMO. Larger studies need to evaluate the safety and efficacy of additional PGE1 in ECMO. Clinical trial registered with EudraCT (2015-005014-30) and www.clinicaltrials.gov (NCT02895373).

Patients on extracorporeal membrane oxygenation are frequently in need for sedation. Use of propofol has been associated with impaired oxygenator function due to adsorption to the membrane as well as lipid load. The aim of our retrospective analysis was to compare two different sedation regimens containing either propofol or midazolam with respect to oxygenator running time. Midazolam was used in 73 patients whereas propofol was used in 49 patients, respectively. In the propofol group, veno-arterial–extracorporeal membrane oxygenation was used predominantly (84%), while veno-venous–extracorporeal membrane oxygenation was used more often in the midazolam group (64%). Oxygenator running time until first exchange was 7 days in both groups (p = 0.759). No statistically significant differences could be observed between the subgroup of patients receiving lipid-free (n = 24) and lipid-containing (n = 31) parenteral nutrition, respectively. Laboratory parameters like triglycerides, free hemoglobin, fibrinogen, platelets, and activated partial thromboplastin time were not significantly different between both sedation regimens (p = 0.462, p = 0.489, p = 0.960, p = 0.134, and p = 0.843) and were not associated with oxygenator running time. The use of propofol as sedative seems suitable in patients undergoing extracorporeal membrane oxygenation therapy.

The duration of extracorporeal membrane oxygenation (ECMO) treatments increases, however, data presented from prolonged support is limited. We retrospectively analyzed all patients during a 4-year period undergoing respiratory ECMO for duration of therapy, demographics, therapy-associated parameters, and outcome according to ECMO duration (<28 days and ≥28 days = long-term ECMO). Out of 55 patients undergoing ECMO for ARDS or during bridging to lung transplantation, 18 were on ECMO for ≥28 days (33%). In the long-term group, median ECMO run time was 40 days (interquartile range 34–54 days). Hospital survival was not significantly different between the groups (54% in short-term and 50% in long-term ECMO patients). There was a significantly higher proportion of patients suffering from malignancy in the group of long-term nonsurvivors. Recovery occurred after more than 40 days on ECMO in 3 patients. The longest ECMO run time in a hospital survivor was 65 days. Duration of ECMO support alone was no prognostic factor and should not represent a basis for decision-making. In patients suffering from malignancy, long-term ECMO support seems to be a factor of adverse prognosis, if not futile.

Background: Monitoring of blood coagulation is essential in ECMO patients. We investigated the prevalence of lupus anticoagulant (LA) and its association with coagulation testing and hemostaseologic complications in patients treated with ECMO. Methods: This is a retrospective analysis including adult patients who received ECMO at a medical intensive care unit at the Medical University of Vienna. The primary outcome was the prevalence of LA. Secondary outcomes included conditions associated with LA positivity, rates of bleeding and thromboembolic events, as well as the proportions of aPTT and antiXa measurements within the target range. Results: Between 2013 and 2021 193 patients received ECMO, in 62 (32%) of whom LA diagnostics were performed. Twenty-two (35%) patients tested positive. LA positive patients had more frequently received VV ECMO (77.3% vs 34.3%; p = 0.002), were more frequently diagnosed with viral respiratory infections (SARS-CoV2: 45.5% vs 20%; p = 0.041, influenza virus: 22.7% vs 0%; p = 0.003), had a longer ECMO treatment duration (25 vs 10 days; p = 0.011) and a longer ICU stay (48 vs 25 days; p = 0.022), but similar rates of bleeding and thromboembolic events.