Preserving the Membrane Lung: An Ongoing ECMO Challenge

To the Editor: We read with interest the article by Mang et al.1 Indeed, oxygenator thrombosis remains a common reason for circuit exchange, the risk of which increases with the duration of extracorporeal membrane oxygenation (ECMO) support.2 Coinciding with the pandemic’s unprecedented increase of ECMO disposable demand, judicious use of limited supply was warranted. While we applaud the authors for their clever approach, we would like to add words of caution regarding their strategy. Catheter-directed regional thrombolysis using a slow thrombolytic infusion is well established, and similarly, the strategy of Mang et al.1 is to deliver the thrombolytic to the site of the thrombus to maximize its local exposure, while minimizing systemic effects.3 This strategy achieves a favorable risk-benefit profile by 1) maximizing thrombus-drug exposure through heightened local drug concentrations; 2) reducing nonlocal side effects as the half-life of alteplase is short (~5 minutes) and agent washout is limited due to vascular occlusion; and 3) limit bleeding complications as compared to bolus thrombolytics application.3 In the absence of detailed thrombolytic dosing rates by Mang et al.1 it is not possible to ascertain whether a sustained extended infusion or a bolus strategy was employed. If the rapid administration of alteplase was applied, it may provide an explanation for why three of the four patients required multiple thrombolytic administrations and one patient receiving nine separate administrations. This presumed bolus administration pattern may increase the risk of devastating bleeding complications. Of additional concern was that patient 1, the lone discharge survivor, had ECMO circuit flow adjusted lower by over 50% throughout the course of eight injections, which may have clouded the pressure drop effects of thrombolysis. This is due to the physical effects of a decreased pressure drop with decreasing flows across the oxygenator. Further, we ask: should the approach to membrane lung preservation rely on rescuing an exhausted device, or rather focus on implementing preventative strategies? Mang et al. reported heparin anticoagulation titrated to an activated partial thromboplastin time of 45–55 seconds, a modest level of anticoagulation. In standard venovenous (VV)-ECMO circuits, this may be sufficient, however, higher anticoagulation levels may be needed in coronavirus disease 2019 (COVID-19) patients due to an underlying hypercoagulable state.4 Furthermore, Mang et al.1 do not report any intensification of anticoagulation in an attempt slow the increasing rate of pressure drop. Finally, one might consider direct thrombin inhibitors as first-line anticoagulants in VV-ECMO patients to mitigate these deposits. Indeed, in aggregate analyses of non-COVID-19 ECMO patients, bivalirudin use was associated with reduced circuit-related thrombotic events when compared to heparin.5 This may be due to pharmacologic differences between the agents. Specifically, heparin only exerts an antithrombotic effect on freely circulating thrombin, whereas bivalirudin exerts its effects on circulating and clot-bound thrombin. This may be advantageous as ECMO circuitry components (oxygenator, connectors, etc.) are prone to fibrin deposition and thrombus prorogation. In our opinion, utilizing bivalirudin as a primary anticoagulant may prolong membrane lung integrity and obviate the need for potentially dangerous rescue maneuvers such as reported by Mang et al.1, at least until the systematic evaluation is available.