Lower Flow, Higher Costs? Recognizing Tradeoffs on the Spectrum of Extracorporeal Support for Acute Respiratory Distress Syndrome

A recent randomized controlled trial (1), a post hoc Bayesian analysis of the same trial (2), and meta-analyses have all suggested a mortality benefit from extracorporeal membrane oxygenation (ECMO) combined with an “ultra–lung-protective” ventilation strategy in patients with very severe acute respiratory distress syndrome (ARDS), presumably through reductions in ventilator-induced lung injury (VILI) beyond what is achieved by the current standard of care (3, 4). More recent emulated target trials during the coronavirus disease (COVID-19) pandemic suggest that benefit may extend to those with Pa_O2:Fi_O2 ratios as high as 120 mm Hg (5, 6). In that context, there has been increasing interest in the use of extracorporeal support to improve outcomes in patients with less severe ARDS through similar reductions in VILI. In patients with relatively well-preserved oxygenation, the focus of extracorporeal support shifts from needing high blood flow rates to support oxygenation to instead ensuring that the device is able to remove enough carbon dioxide to achieve meaningful reductions in the intensity of mechanical ventilation. Indeed, several studies have demonstrated the ability of extracorporeal support at low blood flow rates—where the primary intention is extracorporeal carbon dioxide removal (ECCO₂R)—to facilitate reductions in ventilator parameters well below the current standard of care, to as low as 3 ml/kg predicted body weight, while maintaining pH and Pa_CO2 at similar levels as preextracorporeal support (7–9).

In this issue of the Journal (pp. 1183–1193), Brusatori and colleagues used an experimental lung injury model to investigate the effects of extracorporeal support at high (50–60 ml/kg/min) and low (400 ml/min) blood flow rates—intended to distinguish ECMO from ECCO₂R—on gas exchange, respiratory mechanics, and hemodynamics (10). Their findings, which were influenced by the particular modes of injury and the prescribed changes in invasive mechanical ventilation, all within the initial 24 hours of injury, demonstrated that ECMO resulted in better control of gas exchange, improved hemodynamics (particularly related to improvements in pulmonary vascular resistance), and worse respiratory mechanics, whereas ECCO₂R was associated with worse hemodynamics and the need for higher minute ventilation to achieve adequate carbon dioxide removal (albeit in the setting of higher CO₂ production).

Although the results from the study by Brusatori and colleagues may not be immediately translatable to clinical practice, their results call attention to a key aspect of extracorporeal support as it relates to facilitating ventilator strategies designed to minimize lung injury. The amount and type of extracorporeal support provided to a given patient—like any form of organ support in critical care—should be tailored to meet that particular patient’s needs and must be weighed against its potential risks. In general, we typically strive to provide the least amount of support required to facilitate the modest physiological goal required—a strategy that attempts to minimize the costs associated with many critical care interventions. For example, one of the main tenets of the ventilator strategy and gas exchange goals established by the ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome (ARMA) trial is that there are costs of lung-protective ventilation from minimizing tidal volumes and airway pressures (i.e., hypoxemia and respiratory acidosis), but they are outweighed by the benefits from the reduction in VILI. Extracorporeal support should be thought of similarly; would the patient derive benefit from even greater reductions in mechanical ventilation intensity than the current standard of care, how much extracorporeal support will be needed to achieve those targets, and are the potential complications incurred from extracorporeal support more acceptable than the complications of ongoing mechanical ventilation at higher intensity?

To adequately assess that risk–benefit ratio, it is important to first acknowledge that extracorporeal support exists on a continuum (Figure 1). ECMO and ECCO₂R are not defined by a particular blood flow rate but rather by their intention: ECMO to provide oxygenation (with or without the need for CO₂ removal), and ECCO₂R to provide CO₂ removal (without the intention of providing oxygenation). Therefore, different degrees of extracorporeal support will be needed for different patients (e.g., moderate vs. severe ARDS), as well as over the course of their critical illness (e.g., early vs. later). Whereas low blood flow rates (e.g., <2 L/min) will not meaningfully contribute to oxygenation and would not be defined as ECMO, ECCO₂R efficiency will likewise be affected by very low blood flow rates, as demonstrated in a secondary analysis of the Strategy of Ultra-Protective lung ventilation with Extracorporeal CO₂ Removal for New-Onset moderate to severe Ards (SUPERNOVA) trial, where flow rates of 800–1,000 ml/min were significantly more effective at facilitating tidal volume reductions than rates of 300–500 ml/min (11). Furthermore, the risks incurred by extracorporeal support will vary based on the blood flows applied, devices used, and cointerventions required (e.g., concomitant use of systemic anticoagulation to prevent circuit-related thrombosis). For example, computational and in vitro modeling suggest that, in circuits designed for optimal performance at flows >4 L/min, flow rates below 2 L/min are associated with increased rates of hemolysis and other hematological derangements that may increase bleeding risk (11), which may be further compounded by the potential need for higher levels of anticoagulation to prevent thrombosis at these lower flow rates. In fact, the pRotective vEntilation with veno-venouS lung assisT (REST) trial reported a substantial rate of bleeding, including intracerebral hemorrhage, in the intervention arm (12, 13).

Figure 1.

The continuum of extracorporeal support for acute respiratory failure, characterized by the degree of oxygenation and carbon dioxide removal, the intent of various modes of support, and areas requiring further research. ARDS = acute respiratory distress syndrome; ECCO₂R = extracorporeal carbon dioxide removal; ECMO = extracorporeal membrane oxygenation; MV = mechanical ventilation.

In the end, Brusatori and colleagues demonstrate that similar physiological goals can be achieved by a variety of methods involving some combination of extracorporeal support (be it higher or lower flow) and invasive mechanical ventilation, but several factors need to be better defined to know where the right balance lies. Perhaps patients most likely to derive benefit from reductions in intensity of mechanical ventilation via lower-flow ECCO₂R are best identified by some threshold of respiratory system elastance (14) or ventilatory ratio (15), or maybe only above some minimum Pa_O2:Fi_O2 ratio (15), below which the preferred extracorporeal approach should be higher flow. Defining the optimal “dose” of mechanical ventilation for a given patient will inform both the amount of extracorporeal support needed and whether the tradeoff between extracorporeal and mechanical ventilation risks is acceptable. If low-flow extracorporeal support is associated with unacceptable complications that are insurmountable with current technology, then perhaps mid-to-high flow, more akin to ECMO, should be the preferred approach until we have devices better suited for low flow. Seeing as the relative costs and complications of these strategies remain unclear, the best acceptable compromise requires ongoing evaluation with this type of pilot work before embarking on large-scale clinical trials.