Extracorporeal life support in adults: Management of venovenous extracorporeal membrane oxygenation (V-V ECMO)

INTRODUCTION — Extracorporeal membrane oxygenation (ECMO) is an advanced form of life support that is mostly used in patients with severe respiratory or cardiac failure when conventional treatments fail. While in the past ECMO was associated with poor outcomes and high complication rates, technical advances coupled with accumulating data that describe successful outcomes have resulted in increased ECMO use over time.

This topic will provide an overview of implementation and management of venovenous (V-V) ECMO in patients with acute respiratory failure. Other ECMO-relevant topics are found elsewhere:

●(See "Extracorporeal life support in adults in the intensive care unit: Overview".)

●(See "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)".)

●(See "Extracorporeal life support in adults: Extracorporeal carbon dioxide removal (ECCO2R)".)

●(See "COVID-19: Extracorporeal membrane oxygenation (ECMO)".)

●(See "Extracorporeal life support in adults in the intensive care unit: The role of transesophageal echocardiography (TEE)".)

●(See "Extracorporeal life support in adults in the intensive care unit: Vascular complications".)

TERMINOLOGY — V-V ECMO is a form of extracorporeal life support (ECLS) that is mostly used for temporary support of patients with acute or acute on chronic respiratory failure. The term V-V refers to the configuration of the circuit in which blood is removed from a large central vein (V-V). This blood is oxygenated extracorporeally and reinfused into the same or a different vein (V-V) via a drainage cannula. Additional ECLS terms are described in the table (table 1) and expanded details are discussed in a separate topic. (See "Extracorporeal life support in adults in the intensive care unit: Overview", section on 'Terminology'.)

CLINICAL APPLICATIONS — Our approach to selecting patients for V-V ECMO is discussed in this section. Guidelines that describe the indications and practice of ECLS/ECMO are published by the Extracorporeal Life Support Organization (ELSO) and the American Thoracic Society (ATS) [1-3].

Severe acute respiratory failure — Patients suited to V-V ECMO include patients who have severe, acute, and potentially reversible hypoxemic or mixed hypoxemic and hypercapnic respiratory failure despite optimization of conventional management (eg, lung-protective ventilation [≤6 mL/kg predicted body weight and plateau airway pressure 30 cm H₂O], prone positioning, neuromuscular blockade). The primary goal is oxygenation and CO₂ removal. Examples include those who have unacceptably high airway pressures (eg, plateau pressure >32 cm H₂O) or those who demonstrate intolerance of lung-protective ventilation, such as refractory hypoxemia or severe acidemia [4,5]. These features are mostly encountered in patients with severe acute respiratory distress syndrome (ARDS) but can also be seen in acute respiratory failure due to other conditions. Data to support ECMO in ARDS are derived from randomized trials that suggest benefit while data for other indications are mostly derived from observational studies and clinical experience. These data are discussed in detail below. (See 'Severe ARDS' below and 'Others' below.)

Patients who are not suitable for V-V ECMO include patients who primarily need cardiac support or patients who only need CO₂ removal; however, we may consider them for venoarterial (V-A) ECMO and extracorporeal CO₂ removal (ECCO₂R), respectively. (See "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)" and "Extracorporeal life support in adults: Extracorporeal carbon dioxide removal (ECCO2R)" and 'Assess contraindications and technical challenges' below.)

Although the principles of assessment are similar, data provided in this section are most relevant to patients who have acute respiratory failure that is not due to coronavirus disease 2019 (COVID-19). ECMO efficacy in COVID-19 is discussed separately. (See "COVID-19: Extracorporeal membrane oxygenation (ECMO)", section on 'Indications and contraindications'.)

Severe ARDS — Severe ARDS is the most common indication for V-V ECMO. The criteria that we use and evidence to support them are discussed in this section and shown in the algorithm (algorithm 1).

Criteria for ECMO — Although there are no universally agreed upon criteria for initiation of ECMO in patients with ARDS, we use the entry criteria for a major randomized trial that suggested a likely benefit from ECMO in this population (EOLIA) [6]. This includes patients with severe ARDS who have any one of the following:

●Partial pressure of arterial oxygen to fraction of inspired oxygen [PaO₂:FiO₂] <50 mmHg for >3 hours

●PaO₂:FiO₂ <80 mmHg for >6 hours

●pH <7.25 and partial pressure of arterial CO₂ [PaCO₂] ≥60 mmHg for >6 hours (with respiratory rate at 35 breaths per minute and mechanical ventilation settings adjusted to keep plateau airway pressure ≤32 cm H₂O)

Efficacy — Major randomized trials have suggested benefit associated with ECMO in patients with severe acute respiratory failure, which have mostly included patients with ARDS [6-8]. Most commonly cited are the conventional ventilatory support versus ECMO for severe acute respiratory failure (CESAR) trial and the ECMO for severe ARDS (EOLIA) trial. Despite trial limitations, we believe on balance that ECMO benefits those who meet the above criteria despite optimal conventional management, especially when ECMO is combined with ventilator settings targeting lower volumes and airway pressures than commonly used, and when administered early in the course of ARDS (eg, within the first seven days of invasive mechanical ventilation). (See 'Criteria for ECMO' above and 'Timing and transfer issues' below.)

Data to support ECMO in patients with severe refractory ARDS include the following:

●Meta-analyses – Several meta-analyses that have incorporated CESAR and EOLIA data have reported benefit from ECMO in patients with severe ARDS [3,9-14]. Two meta-analyses reported the 60-day mortality rate was lower in patients receiving V-V ECMO (34 versus 47 percent; risk ratio [RR] 0.73, 95% CI 0.58-0.92) [9,14]. Another meta-analysis reported a reduction in mortality at 90 days (RR 0.75, 95% CI 0.6–0.94) [11]. Adverse events were assessed as likely higher in those receiving ECMO compared with standard mechanical ventilation strategies (eg, hemorrhage).

●CESAR trial – The CESAR trial randomized 180 patients with acute, potentially reversible respiratory failure (mostly ARDS) to ongoing usual care or transfer to a single ECMO-capable center for consideration of ECMO after a period of optimization. The combined outcome, death or severe disability at six months, was significantly lower in the ECMO referral group than the usual care group (37 versus 53 percent; RR 0.69, 95% CI 0.05-0.97) [8]. However, only 70 percent of patients in the conventional group received lung-protective ventilation, and approximately one quarter of those transferred for consideration of ECMO never received it, either because of death or improvement with conventional management. There was also a risk associated with transfer to the ECMO-capable center (ECMO support was not used to facilitate transfer), with three patients dying while awaiting transfer and two others who died during transport. Limitations of this trial included heterogeneous ventilation strategies in the conventional treatment arm, lack of universal application of ECMO in the intervention arm (by design), and a cohort that did not consist solely of patients with ARDS.

●EOLIA trial – The EOLIA trial was the largest trial of ECMO versus standard of care management in severe forms of ARDS [6]. Patients with severe forms of ARDS not responsive to conventional management (see 'Criteria for ECMO' above) were randomized to ongoing optimal conventional measures (lung-protective ventilation using high positive end-expiratory pressure [15], neuromuscular blockade, and prone positioning) or to ECMO combined with a lung-protective ventilation strategy using volumes and airway pressures below what is typically used (also known as "ultra-lung-protective ventilation"). (See 'Invasive mechanical ventilation' below.)

The study was stopped early for interim results that suggested futility [16]. However, a nonsignificant 11 percent absolute difference in 60-day mortality was reported (35 versus 46 percent; RR 0.76, 95% CI 0.55-1.04). A post-hoc Bayesian analysis suggested a strong likelihood of benefit of ECMO over conventional management, although the magnitude of benefit varied depending on the criteria used for the analysis. Limitations include a high rate of crossover from conventional therapies to ECMO and early termination of the trial with inadequate power to detect a statistically significant difference in achieving the prespecified absolute 60-day mortality difference of 20 percent.

Others — We consider V-V ECMO on an individual basis for other patients with acute hypoxemic respiratory failure (with or without hypercapnia) who are unresponsive to conventional care. Examples include:

●Primary graft dysfunction after lung transplantation – For patients with severe primary graft dysfunction (PGD; eg, grade 3 (table 2)) who have refractory hypoxemia despite optimal medical treatment and ventilator settings, V-V ECMO may be successful [17].

However, rather than strict adherence to EOLIA criteria above (see 'Criteria for ECMO' above), we have a lower threshold to initiate ECMO for patients with PGD at the discretion of and in conjunction with the judgment of the lung transplant team. The rationale for this is based upon the potentially greater risk of ventilator-induced lung injury in a recently transplanted allograft. Similar to ARDS patients, we believe outcomes are better if ECMO is initiated early [17-19].

Data to support V-V ECMO in patients with severe PGD are limited to case series.

•One report described survival after V-V ECMO as 82 percent at 30 days, 64 percent at one year, and 49 percent at five years and were better than those achieved with V-A ECMO [20,21].

•In a registry study, among 151 patients with PGD who received V-V or V-A ECMO support, 42 percent survived the hospital stay [22].

•Should patients with PGD survive with ECMO, case series suggest that long-term survival after the first month is similar to that in patients who never received ECMO [23,24].

Identification and management of PGD are discussed separately. (See "Primary lung graft dysfunction", section on 'Refractory hypoxemia'.)

●Patients with reperfusion pulmonary edema and cardiopulmonary failure after pulmonary thromboendarterectomy and patients with diffuse alveolar hemorrhage – Data to support these indications consist of case reports or case series only [25,26]. These indications are discussed separately. (See "Chronic thromboembolic pulmonary hypertension: Pulmonary thromboendarterectomy", section on 'Reperfusion pulmonary edema'.)

●Patients with acute pulmonary embolism with severe gas exchange impairment but preserved cardiac function – Most patients with acute, severe pulmonary embolism present with hemodynamic instability from right ventricular failure, which may in turn warrant the use of V-A ECMO. However, some patients may have severe impairments in gas exchange (eg, refractory hypoxemia) with preserved hemodynamics and cardiac function, making V-V ECMO an appropriate mode of extracorporeal support [27].

●Patients with hypoxemic respiratory failure due to status asthmaticus – Refractory hypoxemia is unusual in this population unless patients have underlying chronic lung disease or concomitant pneumonia. Most patients develop refractory hypercapnic respiratory failure, which is more suited to ECCO₂R, the details of which are discussed separately. (See "Extracorporeal life support in adults: Extracorporeal carbon dioxide removal (ECCO2R)".)

Severe lung disease with ECMO as a bridge to lung transplantation — Select patients with advanced, irreversible underlying lung disease who have acute or acute-on-chronic respiratory failure and require gas exchange support as a bridge to lung transplantation (BTT) may be candidates for V-V ECMO. Because physical deconditioning is considered by many centers to be a strong relative contraindication to lung transplant eligibility, ECMO may also be considered as a means of optimizing physical conditioning via participation in physical therapy in patients with decompensated respiratory failure before the development of significant debility [28-30]. Factors that contribute to making the decision to use V-V ECMO as BTT are described separately. (See "Extracorporeal life support in adults in the intensive care unit: Overview", section on 'ECLS as a bridge to an endpoint'.)

Data to support this indication are limited [31-37]. In an older review of 14 retrospective studies, one-year survival rates among those receiving ECMO as BTT ranged from 50 to 90 percent [37]. However, most patients were mechanically ventilated and had prolonged hospitalizations (three to six weeks), which may not reflect updated practice in many centers. The successful use of ambulatory V-V ECMO with concomitant endotracheal extubation has since been reported, with one cohort study demonstrating ambulation to be a strong independent predictor of successful ECMO as BTT [32].

The optimal timing for initiation of ECLS as BTT has not been determined. We use our clinical judgement regarding timing of and criteria for initiation of ECMO as BTT. Performing ECLS too early may unnecessarily expose the patient to complications of ECLS while delaying initiation may result in the patient no longer being an appropriate transplant candidate despite ECLS support.

General criteria for lung transplant are discussed separately. (See "Lung transplantation: An overview", section on 'Referral for transplant evaluation' and "Lung transplantation: Disease-based choice of procedure", section on 'Disease-based considerations'.)

INITIAL CLINICAL ASSESSMENT

General assessment — As members of an ECMO team, our approach is the following:

●We use clinical history and examination to evaluate the etiology of respiratory failure (eg, acute respiratory distress syndrome [ARDS]).

●We assess the relative contribution of cardiac failure to the patient's presentation, if any. In many cases this is an estimate based upon clinical findings and echocardiography or pulmonary artery catheter characteristics, if available. (See "Approach to diagnosis and evaluation of acute decompensated heart failure in adults".)

●We inquire about and obtain records that may provide a clue to the severity of any underlying illness or comorbidity (eg, prior pulmonary function tests, prior chest imaging, echocardiography, oncologic records for disease stage, calculated BODE index (calculator 1)).

●Using this information, we assess the potential for disease reversibility or targeted intervention (eg, transplantation) and the prognosis from any underlying comorbidity or complications (eg, severity of underlying lung disease, the presence of multiorgan failure). Further details on assessment of reversibility are provided separately. (See "Extracorporeal life support in adults in the intensive care unit: Overview", section on 'ECLS as a bridge to an endpoint'.)

●We also ensure that ECMO is aligned with the patient goals of care by discussing it with the patient (if feasible) and their surrogates/caregivers. We also explain the alternatives including escalating to or continued use of invasive mechanical ventilation or palliation. (See "Communication in the ICU: Holding a meeting with families and caregivers".)

Once the ECMO team makes the decision to proceed, a plan is made for catheter insertion, provided there is no contraindication. (See "Extracorporeal life support in adults in the intensive care unit: Overview", section on 'Staffing (ECMO teams)' and 'Catheter insertion' below and 'Assess contraindications and technical challenges' below.)

Assess contraindications and technical challenges — This assessment includes the following:

●Contraindications – Once the indication is met, we assess potential contraindications to V-V ECMO.

There are few absolute contraindications to V-V ECMO except for a pre-existing condition that is either incompatible with recovery or has no potential for a long-term life-saving therapy (eg, lung transplantation) (table 3) [38]. Examples include severe irreversible neurologic injury (eg, substantial central nervous system hemorrhage that is recent or expanding), severe irreversible lung disease without the potential for transplantation, end-stage malignancy, and irrecoverable multiorgan failure.

Relative contraindications vary from center to center. In general, we consider patients with the following factors as poor candidates for ECMO:

•Uncontrollable bleeding or high risk for bleeding on anticoagulants (eg, severe thrombocytopenia refractory to transfusions)

•Advanced age (eg, >70 years, particularly if significant frailty, malnutrition, or other comorbidities are present)

•Severe immunocompromised status

•Advanced comorbid conditions that would otherwise limit recovery (eg, advanced chronic systolic heart failure or decompensated cirrhosis)

•Limitations in vascular access (eg, severe contractures of the extremities, deep venous thromboses, venous stenoses from chronic indwelling central access catheters)

•Prolonged duration of mechanical ventilation (eg, ≥7 days), particularly those with ARDS who have prolonged exposure to high airway pressures and/or fractions of inspired oxygen

Several ECMO centers consider morbid obesity as a relative contraindication to ECMO (eg, body mass index [BMI] >40 kg/m²). However, we do not consider this a contraindication unless vascular access is notably problematic. Data from the Extracorporeal Life Support Organization (ELSO) registry and the Multicenter ECMObesity Study also support this approach [39,40]. As an example, one retrospective study of 790 patients receiving ECMO for ARDS reported a lower mortality in patients with obesity (BMI ≥30 kg/m²) compared with patients without obesity (24 versus 35 percent; odds ratio 0.63, 95% CI 0.43-0.93) [40].

Neither pregnancy [41] nor acute kidney injury with a need for dialysis are contraindications to ECMO.

●Technical challenges – We assess technical challenges with ECMO implementation. These include assessing vascular access points, bleeding diathesis or contraindications to anticoagulation, immune competency, duration of mechanical ventilation, and need for ambulation in preparation for transplant.

Timing and transfer issues

●Timing – For patients with acute respiratory failure who are potential candidates for V-V ECMO, outcomes are generally better when it is instituted early during the course of a patient's illness (eg, within seven days of invasive mechanical ventilation), although there is no clear cutoff. The data are strongest for patients with ARDS, and practice varies. (See 'Efficacy' above.)

●Transfer – ECMO is not universally available. Therefore, ECMO candidates may need to be transferred from one institution to a specialized center with ECMO capabilities, among other advanced therapies for acute respiratory failure.

In some cases, ECMO support may be initiated at the originating center (either by the originating center staff or by a mobile ECMO transport team from the receiving hospital) prior to transfer while some patients may transfer without initiating ECMO support, depending on clinical stability and ECMO team availability.

In all cases, the potential benefit from ECMO should be weighed against the risk of transfer. Consideration for transfer with or without ECMO should be carefully coordinated between originating and receiving institutions.

Importantly, clinicians should be aware that initiation of ECMO after transfer will be dependent on re-evaluation and further optimization at the receiving facility [8].

CATHETER INSERTION

Venovenous ECMO configurations and sites — There are several configurations for V-V ECMO (figure 1 and figure 2). Choosing among them is at the discretion of the clinical team and institutional practice and may be influenced by anatomic factors, such as the presence of contractures of the extremities or vascular pathology (eg, deep venous thrombosis) and anticipated amount of extracorporeal support needed. Patency of the vessels is confirmed by ultrasound prior to insertion.

Most commonly, we use an initial two-site cannulation approach (figure 1), which is a common default configuration used by many experts. For this configuration, most commonly, the drainage cannula is placed in a femoral vein and threaded up into the inferior vena cava (IVC) until the tip is 1 to 2 cm above the right atrial-IVC junction; the reinfusion cannula is placed in an internal jugular vein (usually the right) with its tip in the superior vena cava (SVC) [42]. Therefore, the tips of the cannulae lie in close proximity to one another but are typically far enough apart to avoid untoward recirculation between the drainage and reinfusion cannulae (see 'Troubleshooting circuit dysfunction' below). This is typically a bedside procedure that does not need fluoroscopy, with cannula placement guided by ultrasonography and positioning confirmed by chest radiography postprocedure.

Other commonly used configurations include the following:

●Dual-lumen, single-site, bicaval cannulae are available that are percutaneously inserted into the internal jugular or subclavian veins with drainage lumens open to both the SVC and IVC, thereby maximizing venous drainage; the reinfusion port is situated in the right atrium with reinfusion flow directed toward the tricuspid valve (figure 2) [43]. This configuration is associated with reduced recirculation compared with a two-site configuration but is technically harder to place, often requiring fluoroscopic and/or transesophageal echocardiographic guidance for accurate placement, and the rate of extracorporeal blood flow is limited by relatively smaller drainage and reinfusion lumens sharing the diameter of the cannula despite an overall larger cannula size.

●A two-site femoral cannulation approach, where the drainage cannula is placed in one femoral vein and threaded up into IVC until the tip is 1 to 2 cm below the right atrial-IVC junction; the reinfusion cannula is placed in the contralateral femoral vein with its tip near but below the right atrial-SVC junction. Maintaining the drainage cannula tip below the reinfusion cannula tip helps minimize recirculation. We consider this approach when internal jugular venous access is technically infeasible.

●Less commonly used is a single-site venous-to-pulmonary arterial (V-PA) dual-lumen cannula (technically classified as venoarterial rather than V-V support), akin to a pulmonary artery catheter in its positioning, that has been proposed for patients with severe acute respiratory distress syndrome (ARDS) with concomitant severe right ventricular dysfunction (figure 3). It is inserted via the internal jugular vein, courses through the right atrium and right ventricle and the tip lies in the pulmonary artery. The drainage port is in the right atrium and the reinfusion port is in the pulmonary artery (image 1). Fluoroscopy is needed for placement.

Size — For two-site V-V ECMO, typical drainage cannula sizes range from 23 to 29 French, with reinfusion cannulae ranging from 18 to 22 French. We select cannula sizes based on the patient's anticipated extracorporeal blood flow needs. For the majority of our patients with ARDS, we use a 29 French drainage cannula and a 22 French reinfusion cannula. Dual-lumen, single-site devices are also available in a range of sizes (up to 32 French).

In V-V ECMO, because of the dependence of oxygenation on the rate of ECMO blood flow (table 4), larger-sized drainage cannulae are typically used for hypoxemic patients, especially those with higher anticipated cardiac outputs (eg, those with a larger body surface area or those in, or at risk for, vasodilatory shock or other high cardiac output states). Preliminary data suggest that larger cannulae may be associated with improved outcomes, although prospective data are needed to validate these findings [44]. However, smaller cannulae may be selected based on lower anticipated needs. While some experts estimate cardiac output based on body size, actual cardiac output may differ substantially from these estimates in critically ill patients. We use judgment based on the overall clinical picture and the patient's anticipated physiologic needs but err on the side of larger drainage cannulae.

Insertion technique — Insertion technique depends upon the cannulae that are inserted.

●For a femoral-internal jugular venous configuration, we use a percutaneous approach via Seldinger technique under ultrasound guidance as the default approach for cannulation [45], although surgical dissection to identify the vessel may occasionally be necessary.

●Single-site cannulae (eg, dual-lumen bicaval or V-PA cannulae) are percutaneously placed but require fluoroscopic or echocardiographic guidance for confirmation of correct placement. The value of echocardiography for proper line placement and monitoring is discussed separately. (See "Extracorporeal life support in adults in the intensive care unit: The role of transesophageal echocardiography (TEE)", section on 'Uses of echocardiography for extracorporeal membrane oxygenation'.)

We secure cannulae with a purse string suture at the cannulation site with additional securement sutures at several points distal to the insertion site along the skin to decrease the risk of displacement and place a sterile dressing over the insertion site. We record the depth of the cannula based on distance between the end of the coil-reinforced portion of the cannula and the insertion site at the skin.

We typically use chest radiography at the end of the procedure to check for appropriate placement and to rule out pneumothorax.

The site should be maintained sterile, with dressing changes as per institutional protocol for maintenance of central venous catheters.

An initial bolus of unfractionated heparin (UFH) is administered just prior to cannula insertion, and a continuous intravenous UFH infusion is started postcannulation once the position is confirmed. Alternative anticoagulants may be considered based on institutional preference or in cases of suspected or confirmed heparin-induced thrombocytopenia. (See 'Anticoagulation' below and "Management of heparin-induced thrombocytopenia".)

INITIAL SETTINGS AND TITRATION — Following cannulation, the patient is immediately connected to the ECMO circuit (typically the pump, membrane lung, and console) and the sweep gas (a blend of oxygen and air, the fractions of which are determined by the provider and set through a blender) is connected to the membrane lung.

There are no universally agreed-upon initial settings. Typical settings are dictated by the degree of support that may be needed. As an estimate we typically use the following, achieved incrementally over the first few minutes after initiation:

●Pump speed is increased to achieve an extracorporeal blood flow of at least 3 L/minute (range 3 to 7 L/minute).

●Sweep gas flow typically ranges from 1 to 10 L/minute, with initial settings usually at the lower end of that range. Sweep gas flow may be titrated based on pH to offset a component of respiratory acidosis or in order to decrease respiratory drive in some patients.

●Fraction of delivered oxygen of 1.0.

Immediately following initiation, we perform the following:

●The blood flow rate is increased until oxygenation is satisfactory such as peripheral arterial oxygen saturation (SpO₂) >87 percent, although, in general, the goal for oxygenation should be adequate delivery of oxygen to end-organs (eg, decreasing blood lactate levels, decreasing creatinine, increasing urine output) without focusing on a specific target SpO₂. (See "Measures of oxygenation and mechanisms of hypoxemia", section on 'Commonly used'.)

●Sweep gas flow is titrated often with a target of near-normal pH (if there is not a substantial concomitant metabolic acidosis) or, in select patients, reduced respiratory drive. Of note, for those with marked, uncompensated respiratory acidosis (eg, arterial carbon dioxide tension [PaCO₂] >60 mmHg with corresponding acidemia), the sweep gas flow rate should be gradually increased over several hours to avoid complications associated with rapid changes in PaCO₂ (eg, intracerebral hemorrhage) [46,47]. The precise rate of change in PaCO₂ that may result in harm is unknown. However, the more rapid the increase in PaCO₂ prior to ECMO, the more rapidly the PaCO₂ may be safely lowered.

Within an hour of cannulation, we generally obtain routine laboratories, systemic arterial blood gas (ABG; eg, radial ABG), coagulation studies, and pre- and post-membrane lung blood gases.

In parallel with titration of ECMO settings, we begin to adjust mechanical ventilation settings. Further details are provided below. (See 'Invasive mechanical ventilation' below.)

Daily assessment and circuit monitoring are discussed separately. (See 'Daily assessment and circuit monitoring' below.)

SUPPORTIVE THERAPIES

Invasive mechanical ventilation — For patients receiving V-V ECMO, there is no universally agreed upon strategy for ventilator management, with practice varying among clinicians and centers. However, experts generally agree on using a lung-protective strategy that minimizes ventilator-induced lung injury (VILI) beyond what can typically be achieved with conventional mechanical ventilation alone. Data to support such a lung-protective approach is stronger for patients with acute respiratory distress syndrome (ARDS) compared with those mechanically ventilated for other reasons.

●Patients with ARDS – In our center, for patients with ARDS (table 5), we typically use an "ultra-lung-protective" ventilator strategy (ie, tidal volumes and airway pressures below conventional practice in ARDS without the use of ECMO). The rationale for this strategy is based on criteria in the EOLIA trial that suggested benefit when V-V ECMO was combined with an "ultra-protective" strategy in patients with severe ARDS [6] (see 'Efficacy' above). Components of this strategy are listed in the table (table 5).

In general, we make the specified adjustments to the tidal volume, respiratory rate, positive end-expiratory pressure (PEEP), and fraction of inspired oxygen in tandem with ECMO titration immediately after cannulation. When patients are cannulated at referring hospitals and must be transported to their destination institution, we make more gradual initial adjustments to PEEP and tidal volumes (over a few hours); this avoids alveolar derecruitment during transport while still minimizing exposure to excessively high airway pressures and tidal volumes.

●Patients without ARDS – In most patients without ARDS who require V-V ECMO, we use a similar strategy to that outlined for ARDS in the above bullet, although there are very limited data in this population to support this practice.

●Bridge to lung transplantation – For patients receiving ECMO as bridge to lung transplantation (BTT), we attempt to completely liberate patients from invasive mechanical ventilation (ie, endotracheal extubation) if feasible. The rationale is that this is achievable in many patients when ECMO is sufficient for gas exchange and mechanical ventilatory support is no longer needed for work of breathing. In such situations, this may facilitate the patient's participation in physical therapy in preparation for lung transplantation in order to optimize and maintain transplant candidacy. This strategy can also be occasionally performed in some patients with severe ARDS prior to decannulation from ECMO, although endotracheal extubation and early mobilization have yet to be associated with improved outcomes in patients with ARDS receiving ECMO support [30]. Ventilator weaning and extubation strategies are discussed separately. (See "Weaning from mechanical ventilation: Readiness testing" and "Initial weaning strategy in mechanically ventilated adults" and "Extubation management in the adult intensive care unit".)

Anticoagulation — Due to the high risk of thrombosis, patients receiving V-V ECMO should be anticoagulated in order to maintain circuit patency, provided there is no contraindication [48,49]. However, the agent used, target level of anticoagulation, and monitoring varies from center to center [50]. In our center, we administer an initial bolus of unfractionated heparin (UFH; eg, 5000 units) just prior to cannula insertion and start a continuous intravenous UFH infusion postcannulation. We use a low-level anticoagulation strategy targeting an activated partial thromboplastin time (aPTT) of approximately 40 to 60 seconds, although aPTT values may vary from center to center and should reflect local practice. The International Society of Thrombosis and Hemostasis have issued guidance on this matter [49].

These anticoagulation targets may be superseded by other indications that warrant higher levels of anticoagulation (eg, venous thromboembolism, atrial fibrillation, thrombosis in other sites) or by the need to temporarily suspend the use of anticoagulation in the setting of bleeding or a procedure.

In circumstances where anticoagulation is temporarily contraindicated, we maintain ECMO blood flow rates >3 L/minute to minimize hemostasis or hemolysis (especially within the pump head) and closely monitor the patient for evidence of circuit thrombosis that might necessitate circuit exchange. When feasible, we use an anticoagulant at prophylactic doses until the risk for bleeding normalizes and therapeutic anticoagulation can be resumed. (See 'Troubleshooting circuit dysfunction' below.)

Some centers use activated clotting time, factor Xa levels, or thromboelastography as monitoring tools for optimal anticoagulation levels. Similarly, although performed in some centers, we do not routinely measure or replete antithrombin III. The International Society of Thrombosis and Hemostasis have issued guidance on potential targets for these parameters [49].

Transfusion — For patients receiving ECMO for ARDS, we use similar red blood cell transfusion thresholds to those in the general critically ill population (eg, hemoglobin below 7 g/dL) [51]. This practice is supported by a prospective cohort study of adults receiving ECMO support for ARDS, in which transfusion of packed red blood cells was associated with reduced mortality only when the pretransfusion hemoglobin was below 7 g/dL [52]. This transfusion strategy, when combined with low anticoagulation targets and auto-transfusion of circuit blood at the time of decannulation, has been associated with minimal transfusion requirements and favorable overall survival [48]. However, in transplant candidates, the transfusion threshold may be set at a lower hemoglobin level to avoid allosensitization, with a transfusion trigger based on end-organ function (as a surrogate for adequate oxygen delivery) and exercise tolerance. (See "Use of blood products in the critically ill", section on 'Restrictive strategy as the preferred approach'.)

The transfusion of platelets and fresh frozen plasma should also follow usual critical care thresholds (eg, platelet counts <20,000/microL or <50,000/microL in the presence of active bleeding or procedures). (See "Use of blood products in the critically ill", section on 'Plasma' and "Use of blood products in the critically ill", section on 'Platelets'.)

Routine measures — Other supportive measures may have ECMO-specific issues but are mostly discussed in the linked topics below:

●Fluid and electrolyte management – As for most patients with ARDS, we prefer conservative fluid management with buffered or nonbuffered crystalloids, with diuresis as needed to manage pulmonary edema. Such a strategy has been associated with a lower 90-day mortality in a large cohort of patients receiving ECMO for refractory respiratory or cardiac failure [53]. However, other patients and ARDS patients with other conditions may require a more liberal strategy (eg, ARDS plus vasodilatory shock and end-organ hypoperfusion, ARDS due to pancreatitis). Further details are provided separately. (See "Treatment of severe hypovolemia or hypovolemic shock in adults" and "Acute respiratory distress syndrome: Fluid management, pharmacotherapy, and supportive care in adults", section on 'Conservative fluid management' and "Evaluation and management of suspected sepsis and septic shock in adults", section on 'Intravenous fluids (first three hours)'.)

●Sedation, analgesia, mobilization – In patients receiving ECMO, higher doses of sedatives and analgesics may be necessary to achieve the same sedation targets as patients not receiving ECMO; this may be due to both sequestration of more lipophilic and protein-bound medications within the circuit and higher volumes of distribution [54].

Where some patients receiving ECMO may only tolerate ultra-low tidal volume strategies with deeper sedation targets, others may be able to tolerate these settings with lower sedation targets combined with higher sweep gas flow rates that may blunt respiratory drive [55,56].

In some patients, sedation may be weaned off successfully when the goal is early mobilization for patients who need V-V ECMO as a BTT and, rarely, in select patients with ARDS who can tolerate extubation prior to decannulation [30,31,57,58]. (See 'Invasive mechanical ventilation' above.)

Early physical therapy, including mobilization out of bed, has been demonstrated to be safe and feasible in cohort studies and a pilot randomized trial [57-60]. In a large observational study of patients with cardiopulmonary failure, V-V ECMO with BTT as the indication for ECMO was associated with greater odds of achieving out-of-bed versus in-bed activities compared with V-A ECMO [59]; femoral cannulation and concomitant need for invasive mechanical ventilation were associated with decreased odds of performing out-of-bed activities.

Further details are provided separately. (See "Sedative-analgesia in ventilated adults: Management strategies, agent selection, monitoring, and withdrawal" and "Post-intensive care syndrome (PICS) in adults: Clinical features and diagnostic evaluation".)

●Tracheostomy – The indications for tracheostomy are similar to patients not receiving ECMO, but the procedure may carry a higher risk in those receiving ECMO. In an retrospective study, patients undergoing tracheostomy during ECMO support had a higher rate of procedural site bleeding compared with those receiving tracheostomy after ECMO decannulation [61]. Timing of tracheostomy, including whether to perform it during or after ECMO support, should be considered on a case-by-case basis, taking into consideration the anticipated duration of ECMO and invasive mechanical ventilation. Further details are provided separately (See "Tracheostomy: Rationale, indications, and contraindications".)

●Antimicrobials – Similar to sedatives and analgesics, achieving therapeutic levels of antimicrobial medications may be affected by both sequestration of more lipophilic and protein-bound medications within the circuit and higher volumes of distribution. We check drug levels as appropriate and consult with an expert pharmacist for complex cases.

●Ventilator-associated pneumonia precautions – We do not use prophylactic antibiotics, although preliminary retrospective data may suggest a possible benefit [62]. (See "Risk factors and prevention of hospital-acquired and ventilator-associated pneumonia in adults".)

●Nutritional support – (See "Nutrition support in intubated critically ill adult patients: Initial evaluation and prescription" and "Nutrition support in intubated critically ill adult patients: Enteral nutrition" and "Nutrition support in intubated critically ill adult patients: Parenteral nutrition".)

●Glucose control – (See "Glycemic control in critically ill adult and pediatric patients".)

●Stress ulcer prophylaxis – (See "Stress ulcers in the intensive care unit: Diagnosis, management, and prevention" and "Management of stress ulcers".)

●Hemodynamic monitoring – (See "Pulmonary artery catheterization: Indications, contraindications, and complications in adults" and "Pulmonary artery catheterization: Interpretation of hemodynamic values and waveforms in adults" and "Novel tools for hemodynamic monitoring in critically ill patients with shock".)

●Fever management – (See "Fever in the intensive care unit", section on 'Outcomes'.)

Prone positioning — We consider prone positioning as a mechanism that facilitates additional lung protection during V-V ECMO, when needed, but we do not advocate its routine use.

Prone positioning during V-V ECMO has been shown in observational trials to improve respiratory mechanics, oxygenation, and possibly mortality in patients with acute respiratory distress syndrome (ARDS) [63-65]. However, a randomized trial suggested no benefit from the routine application of prone positioning during V-V ECMO for ARDS patients in whom an "ultra-protective" invasive mechanical ventilation strategy (similar to EOLIA) was employed [66]. In a multicenter randomized trial of 170 patients with severe ARDS who were receiving V-V ECMO for less than 48 hours, routine early use of prone positioning (a minimum of four sessions for at least 16 hours each) was associated with similar outcomes as supine positioning. Reported outcomes included rate of successful ECMO weaning at 60 days (44 percent each), ECMO duration (28 versus 32 days), intensive care unit length of stay, and 90-day mortality (51 versus 48 percent). Adverse events were similar among the groups. Cardiac arrest was greater in the supine group, but event rates were low, decreasing confidence in the validity of this outcome. Limitations that may have impacted the results include the large proportion of patients who had received proning before enrollment (>90 percent), the high proportion of COVID-19 ARDS patients (94 percent), and a lack of blinding. Further trials are needed to determine the impact of prone positioning, if any, during V-V ECMO in non-COVID-associated ARDS.

Proning is not typically limited by cannula configuration. Further information on prone ventilation in ARDS patients not receiving V-V ECMO is provided separately. (See "Prone ventilation for adult patients with acute respiratory distress syndrome".)

DAILY ASSESSMENT AND CIRCUIT MONITORING

Circuit management and monitoring — To ensure a properly functioning circuit and detect complications early, we examine patients daily and periodically throughout the day. We obtain the following daily laboratories and imaging:

●Complete blood count

●Chemistries including lactate dehydrogenase, bilirubin, aspartate transaminase (specifically for hemolysis)

●Systemic arterial blood gas

●Coagulation studies per institutional protocol

●Chest radiography

Plasma-free hemoglobin, urinalysis, and peripheral blood smear may also be monitored routinely or reserved for when there is particular concern for hemolysis based on hemoglobin trend, chemistries, and urine appearance or urinalysis. We use a plasma-free hemoglobin threshold of 50 mg/dL as a marker of significant hemolysis that should prompt consideration of a circuit exchange [67].

ECMO circuits are continuously monitored to ensure all components are functioning properly by obtaining the following:

●Pre- (P1) and post- (P2) membrane reinfusion pressures; in general, we aim to maintain line pressures less than 300 mmHg

●Drainage pressure (prepump), with the general goal of not exceeding -100 mmHg to minimize blood trauma

●P1 and P2 membrane blood gases when there is concern for inadequate membrane lung function

●Cannulae position at least daily (eg, chest radiograph, measurement at the skin) and suture integrity

●Evidence of infection and bleeding

●Visible signs of thrombus within the membrane, cannulae, and tubing

All medications, including parenteral nutrition, and the provision of renal replacement therapy are administered through separate intravenous access points whenever feasible.

Troubleshooting circuit dysfunction — During the patient's ECMO course, problems may arise with the circuitry and membrane lung that require immediate attention.

●Membrane lung dysfunction – Membrane lung dysfunction or failure should be suspected when any one or more of the following are present:

•Inadequate systemic oxygenation not otherwise explained by the patient's underlying disease.

•Post-membrane oxygen tension <200 mmHg and oxygen transfer <150 mL/minute at optimized ECMO blood flow and with fraction of delivered oxygen of 1.0; oxygen transfer is circuit blood flow x (CpostO₂ – CpreO₂) x 10, where C is content and x represents whatever source the blood is coming from (eg, 'pre,' 'post,' 'arterial') [68].

•Inadequate ventilation despite high sweep gas flow rate (eg, >10 L/minute).

•Concern for circuit related thrombosis as evidenced by the following:

-Rapidly rising delta membrane gradient (ie, P1 pressure minus P2 pressure) normalized for blood flow rate or not explained by an increase in extracorporeal blood flow.

-A delta membrane gradient >60 mmHg.

-Evidence of blood flow obstruction (eg, increase in revolutions per minute [RPMs] without an accompanied increase in blood flow, decrease in blood flow without a change in RPMs not otherwise explained by excessively negative drainage pressures).

•Evidence of circuit-related coagulopathy (eg, disseminated intravascular coagulopathy, acquired von Willebrand disease, antithrombin III deficiency) as a result of shear forces within the circuit.

•Evidence of hemolysis (eg, decreasing hemoglobin with increased markers of hemolysis, including plasma free hemoglobin >80 mg/dL), not otherwise explained by patient's underlying condition or other etiology.

In most cases of membrane dysfunction, we exchange the membrane. If thrombosis is isolated to the pump head, then that component will require exchange, which, depending on circuit design, may require simultaneous change-out of the membrane lung if they are bound together.

If circuit blood flow needs to be interrupted to exchange one or more components, the clinical team should be prepared to manage the patient during the exchange with higher levels of ventilator support as needed (eg, increasing respiratory rate, tidal volume, fraction of inspired oxygen, and PEEP), although the effectiveness of this depends on the gas exchange abilities of the native lungs. This is referred to as "emergency circuit clamping measures."

●Recirculation – Recirculation is a phenomenon in which reinfused oxygenated blood is drawn back into the circuit without passing through the systemic circulation, thereby decreasing the contribution of ECMO to systemic oxygenation [69]. This phenomenon is more commonly and more strongly associated with two-site configurations (figure 1).

Recirculation is assessed by monitoring the oxygen saturation of blood entering the membrane (SpreO₂). There is typically always some degree of recirculation, and this only warrants intervention if it is having a significant impact on systemic oxygenation (eg, inadequate systemic oxygenation and SpreO₂ approaching or exceeding arterial oxygen saturation).

Typically, it is the result of drainage and reinfusion cannulae being too close to each other, which can be assessed by chest radiography, and is usually managed by retracting one of the cannulae (typically the drainage cannula) by 1 to 2 cm or more if appropriate. Prior to actual retraction of the cannula, gently pulling back on the cannula while it is still sutured to the skin may mimic actual retraction and offer a sense of whether recirculation would improve (as evidenced by a decrease in SpreO₂ and an improvement in systemic oxygenation). Other causes of recirculation include high blood flow rates with excess drainage pressure and increases in intrathoracic pressures.

Initial placement of a single-site, dual-lumen cannula (figure 2) may minimize recirculation when properly positioned [45,70].

●Chatter (or chugging) – Chatter occurs when the drainage pressure is too high relative to the available blood in the draining vessel, resulting in lower, often fluctuating blood flow rates.

Chatter is caused by the vessel wall collapsing around and occluding ports in the drainage cannula or kinking of the tubing in the drainage line.

-Suspected port occlusion – Our preferred approach to managing drainage port occlusion is to reduce the pump speed to decrease the drainage line pressure, so long as the patient can tolerate the resultant lowered blood flow rates (which may, seemingly paradoxically, be higher than it was at the higher RPMs where chatter was occurring).

Alternatively, in patients with suspected intravascular volume depletion, intravenous fluids may be administered to increase blood volume to improve drainage. Such an approach is counterproductive for patients in whom fluid overload or acute respiratory distress syndrome are contributing to their need for ECMO.

-Kinks, if present, should be identified and relieved.

●Circuit air – Entrainment of air may occur anywhere along the portion of the circuit under negative pressure (prepump) where there is loss of circuit integrity or through other venous access points (peripheral or central intravenous lines) that are left open or damaged. A less common mechanism through which gas may enter the circuit is known as cavitation, which occurs when a high negative pressure within the drainage cannula or tubing draws a large amount of gas out of solution, or in the setting of a high sweep gas flow rate and a low blood flow rate. In general, cavitation can be minimized by avoiding extremes in negative pressure on the venous side of the circuit.

Once entrained, air may pass into the membrane lung, which serves as a gas trap. However, large amounts of air may, on occasion, pass through the membrane lung and be delivered to the patient, resulting in air embolism.

When air is identified in the circuit, we urgently inspect for defects in the cannulae, tubing, and access points (including other patient lines and devices, such as dialysis or plasmapheresis cannulae) and check drainage cannulae for exposed drainage ports.

•For pre-membrane air, we isolate the air within the tubing at the closest access point, aspirate it with a syringe, restore circuit volume equal to what was removed with the syringe, and repair or replace any circuit defects.

•For air in the membrane itself, we remove the membrane from its holder, tap the oxygenator to isolate air superiorly, aspirate the air through a syringe, and restore circuit volume. If there is a large amount of air, the oxygenator or circuit may need to be reprimed or replaced.

•If air is detected post-membrane, in order to prevent air embolism to the patient, we clamp the arterial and venous limbs of the circuit, reduce pump speed, follow emergency circuit clamping measures (ie, supporting the patient with maximal conventional management), isolate air within the tubing at the closest access point, aspirate it through a large syringe, and replace blood volume removed during aspiration.

●Tubing disconnection/tubing rupture/cannula fracture – Should tubing accidentally disconnect or rupture or should the cannula fracture (resulting in air entrainment or bleeding), we immediately identify the site and reconnect the tubing if disconnected. If rupture or fracture has occurred, we clamp all limbs of the circuit to isolate the patient from ECMO, reduce the RPMs, follow emergency circuit clamping procedures (supporting the patient with maximal conventional management), replace the tubing at the site of the defect, and transfuse for any clinically relevant blood loss. Once the defect is resolved, the circuit may be unclamped and pump speed increased to where they were prior to the event. If the fracture occurs within the cannula, a new cannula will need to be placed at an alternate access point to restore circuit integrity.

●Heater-cooler malfunction – Signs that there has been a malfunction of the heater-cooler include water in the membrane lung or blood in the heater-cooler's water bath or rapid decrease in the blood temperature. Management involves replacing the heater-cooler, replacing the membrane lung if there is evidence of water or blood leakage, and managing hypothermia (if undesired) through conventional means until the heater-cooler is replaced.

●Inadvertent decannulation – Inadvertent decannulation is a rare but potentially life-threatening event during ECMO support.

If there is partial dislodgement with preserved blood flow, we assess the cannula for exposure of drainage ports. If any drainage ports are exposed, the cannula must be replaced because of the infectious risk. If drainage ports are not exposed and blood flow and line pressures are acceptable, we reinforce the cannula with sutures and obtain a chest radiograph to confirm cannula location.

If there is complete dislodgement or partial dislodgement with compromised blood flow, we clamp all limbs of the circuit to isolate the patient from the circuit, reduce pump speed, follow emergency circuit clamping measures, control cannula site bleeding, discontinue anticoagulation, replace the cannula with a new cannula (assuming that continued ECMO support is still required), assess blood flow and line pressures, and obtain chest radiograph to confirm cannula placement.

●Oxygen source disconnect – Disconnection of the oxygen source from the blender, or the blender from the membrane, may manifest as patient hypoxemia with no difference between the P1 and P2 oxygen content (also evidenced by dark, venous-appearing blood entering and leaving the membrane) with unchanged P1-P2 pressure gradient or other signs of membrane lung dysfunction. In such cases, we reconnect the oxygen source.

●Pump malfunction/failure – Pump failure is an uncommon complication of ECLS. If pump failure occurs, emergency circuit clamping measures should be followed immediately (clamping both drainage and reinfusion lines). Where available, a hand crank or back up console should be kept alongside ECLS circuits and used to generate blood flow (after unclamping all lines) while the source of the problem is identified and corrected.

If pump failure is caused by disruption of the electrical power supply, then battery power or a backup power supply should be obtained.

If the ECLS circuit was inadvertently powered off, it should be restarted immediately, and alarm functionality assessed.

If the circuit has adequate power supply and the pump is rotating, but there is no flow, there may be inadequate preload or excess afterload. Centrifugal pumps may decouple from the motor, in which case the pump speed may be temporarily turned to zero to allow recoupling; if unsuccessful, the pump should be stopped and removed from its seat on the motor. Pump integrity should be confirmed through visual inspection, assessing for cracks or thromboses. If the pump head is intact, it may be placed back onto the motor and the pump restarted. In the event there is still no flow, the motor should be replaced.

COMPLICATIONS — The complications of V-V ECMO are listed in the table (table 6).

Many of the listed complications in the sections below occur with both V-V and venoarterial (V-A) ECMO. Additional V-A-specific ECMO complications are discussed separately. (See "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)", section on 'Other complications'.)

Bleeding — Bleeding is a common complication during ECMO support, the rates of which are site-specific and can be life-threatening [6,8,50,71]. Cannula site and surgical site bleeding are as high as 8 percent while gastrointestinal, pulmonary, or retroperitoneal hemorrhage are less common.

Bleeding is most often in the setting of the administration of continuous anticoagulation and is more often driven by circuit-related or patient-driven platelet dysfunction (eg, acquired von Willebrand syndrome [72]) or thrombocytopenia. Conditions such as disseminated intravascular coagulation (DIC) and vessel or cardiac perforation may also contribute to bleeding rates.

Meticulous surgical technique, maintaining platelet counts >20,000/microL (or >50,000/microL for procedures), and targeting a conservative activated partial thromboplastin time (aPTT; eg, approximately 40 to 60 seconds) reduce the incidence of bleeding.

Most instances of bleeding are minor and may resolve with conservative management (eg, additional suture, manual pressure). For most major bleeding events (eg, bleeding requiring more than two units of packed red blood cells, hemodynamic instability, life-threatening organ involvement), we temporarily reduce or hold anticoagulation for a few hours (or longer, as needed) until the bleeding has stopped. Bleeding from surgical wounds or into accessible body cavities may warrant prompt surgical exploration with use of electrocautery as needed to achieve hemostasis. Under rare circumstances, when the risk of death from bleeding is so high that it outweighs the risk of circuit thrombosis, plasminogen inhibitors (eg, aminocaproic or tranexamic acid) can be infused [73-75]. Infusion of activated factor VII has been reported with mixed results and we only consider it for life-threatening hemorrhage after all other options have failed [76-78].

Hemolysis — Hemolysis may result from shear forces within the circuit, often a result of blood passing through areas of high resistance caused by thrombus or fibrin deposition within the circuit (eg, membrane lung, pump head, cannulae), excessive negative or positive pressure within the circuit, or turbulent blood flow through the pump head at lower blood flow rates (eg, <2 L/minute).

●If clinically significant hemolysis is attributable to a circuit component, then that component may need to be replaced. (See 'Circuit management and monitoring' above and 'Troubleshooting circuit dysfunction' above.)

●Underlying causes of excessive negative pressures (eg, pump speed too high relative to available blood volume in the draining vessel, drainage cannula or tubing occlusion) or excessive positive pressures (eg, occlusion of reinfusion cannula or tubing) should be investigated and addressed when possible.

●The loss of pump head efficiency, increased risk of hemolysis, and other hemocompatibility issues at lower blood flow rates is why we recommend maintaining ECMO blood flow rates >3 L/minute [79-81].

Clinical signs of hemolysis are discussed separately. (See "Non-immune (Coombs-negative) hemolytic anemias in adults" and "Diagnosis of hemolytic anemia in adults", section on 'Laboratory confirmation of hemolysis'.)

Thrombocytopenia — Thrombocytopenia is a common occurrence in patients receiving ECMO.

Thrombocytopenia may be attributable to patients' underlying critical illness, sequestration within the ECMO circuit, consumption in the setting of bleeding, heparin-induced thrombocytopenia (HIT), or DIC. A large single-center cohort study identified baseline severity of illness and pre-ECMO platelet counts, rather than duration of ECMO support, as predictors of severe thrombocytopenia [82].

For patients who develop thrombocytopenia while receiving ECMO, we typically evaluate for associated etiologies and maintain platelet counts >20,000/microL (or >50,000/microL with active bleeding or procedures). Management of HIT and DIC are discussed separately. (See "Management of heparin-induced thrombocytopenia" and "Evaluation and management of disseminated intravascular coagulation (DIC) in adults".)

The diagnostic investigation of thrombocytopenia is discussed separately. (See "Diagnostic approach to thrombocytopenia in adults".)

Thrombosis and thromboembolism — Thrombus formation with or without embolization is a known complication of ECMO that is more likely to occur when anticoagulation is not being administered, although thrombosis may still occur in the presence of systemic anticoagulation. Maintaining targeted aPTT and vigilant observation of the circuit for signs of clot formation, with increases in anticoagulation targets as tolerated, help minimize thromboembolism in most patients.

●Systemic venous thrombosis – Systemic venous thromboembolism is not unusual with some reports suggesting rates of pulmonary embolism (PE) as high as 16 percent [50,83,84]; rates of deep venous thrombosis (DVT) may be higher and may be associated with the presence of cannulae, especially those placed in the femoral vasculature.

Our practice is to systematically evaluate patients for DVTs post-decannulation in all four extremities, regardless of which veins were used for cannulation, although this practice is not evidence-based. However, other groups evaluate only the vessels that were used for cannulation or do not evaluate for thrombosis at all. Strategies used to treat lower extremity DVT and PE are discussed separately. (See "Overview of the treatment of proximal and distal lower extremity deep vein thrombosis (DVT)" and "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults".)

●Circuit thrombosis – We routinely observe the circuit for signs of clot formation, including routine visual inspection of all cannulae, tubing, and connectors, and monitoring the pressure gradient across the membrane lung. An increase in the pressure gradient across the membrane lung (eg, >60 mmHg), when normalized for blood flow rate, raises suspicion for substantial clot within the membrane. Large or mobile clots require immediate circuit or component exchange. (See 'Troubleshooting circuit dysfunction' above.)

●Arterial thrombosis – Systemic arterial and cardiac thrombosis and thromboembolism are more commonly associated with V-A ECMO than V-V ECMO but may occur during V-V ECMO in the presence of venous thrombosis and concomitant intracardiac defects. (See "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)", section on 'Cardiac or aortic thrombosis' and "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)", section on 'Neurologic injury' and "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)", section on 'Peripheral arterial complications'.)

Cannula-insertion-related vascular complications — A variety of rare complications may occur during venous cannulation (<5 percent). These include vessel perforation with hemorrhage and cannula misplacement (eg, venous cannula within the artery), with potential for arterial dissection if the artery is inadvertently cannulated. Cannulation performed by an experienced operator with appropriate surgical backup helps to both minimize and address such complications. (See "Extracorporeal life support in adults in the intensive care unit: Vascular complications".)

Neurologic issues — Neurologic events, such as ischemic stroke and intracranial hemorrhage, have been reported with both V-V and V-A ECMO [46,47,85]. The incidence of neurologic injury in adult respiratory failure patients recorded in the Extracorporeal Life Support Organization (ELSO) registry is approximately 10 percent, with hemorrhagic stroke being more common than infarction [86].

Whether these events are ECMO-related is unclear since neurologic events may also be due to underlying illness. As an example, one randomized trial reported a lower proportion of ischemic and hemorrhagic strokes in those treated with ECMO compared with patients treated with conventional care (0 versus 5 percent) [6].

A common finding in some studies is an association between rapid change in the arterial partial pressure of carbon dioxide (PaCO₂) upon ECMO initiation and intracranial bleeding [46,47]. These data are the basis for the recommendation to slowly correct PaCO₂ over the course of hours after the initiation of ECMO in patients with uncompensated hypercapnia. (See 'Initial settings and titration' above.)

Evaluation and management of stroke and intracranial hemorrhage are discussed separately. (See "Overview of the evaluation of stroke" and "Spontaneous intracerebral hemorrhage: Acute treatment and prognosis" and "Reversal of anticoagulation in intracranial hemorrhage".)

Infections — In patients receiving ECMO, local and systemic infections are common and may be cannula-related or due to the underlying illness. Cannula-associated infections are managed with appropriate antibiotics, but unlike other central line-associated infections, removing or exchanging the cannulae is typically not feasible or advisable except under the most extreme circumstances of severe, refractory infection with associated septic shock that is attributable to the cannulae/circuit. We typically treat through systemic infections. For cannula site infections, washout or cannula replacement may be required.

REFRACTORY HYPOXEMIA OR CONCOMITANT CARDIAC FAILURE — Patients may experience severe, refractory hypoxemia despite V-V ECMO and/or may develop cardiac failure, both of which require management.

Management of refractory hypoxemia during ECMO — For patients on ECMO, we define refractory hypoxemia as that evidenced by end-organ dysfunction attributable to inadequate tissue oxygenation (eg, low systemic oxygen saturation and elevated lactate not otherwise attributable to an alternative cause).

We use the following approach when managing refractory hypoxemia during V-V ECMO:

●Maximize ECMO blood flow rate – In most cases, higher ECMO blood flow rates should mitigate hypoxemia. However, if blood flow is limited by excessively negative drainage pressures or ECMO flow is already maximized, we typically add a second venous drainage cannula (eg, VV-V) to help achieve higher blood flow rates at less negative drainage pressures [87]. This assumes that adequate sedation and ventilator synchrony (with or without the use of neuromuscular blockade) have been achieved.

●Other options – Should the above fail, other options include the following, although occasionally, they are performed before the insertion of an additional drainage cannula, depending on the clinical situation:

•Neuromuscular blockade or targeted temperature management (eg, cooling to 36°C, managed through the circuit's heater-cooler) may minimize oxygen demand [88]. (See "Neuromuscular blocking agents in critically ill patients: Use, agent selection, administration, and adverse effects" and "Intensive care unit management of the intubated post-cardiac arrest adult patient", section on 'Active temperature control'.)

•We rarely use packed red blood cell transfusions to increase oxygen carrying capacity in notably anemic patients, although the risk-to-benefit ratio of this practice is unknown. (See "Use of blood products in the critically ill", section on 'Red blood cells'.)

•It has been proposed that beta blockers be used to decrease cardiac output and thereby increase the proportion of blood flow through the ECMO circuit relative to cardiac output. However, such an approach has been associated with decreased oxygen delivery [89], and we would only consider it when the impact on oxygen delivery is actively being measured. (See "Oxygen delivery and consumption", section on 'Oxygen delivery'.)

•Inhaled pulmonary vasodilators (eg, inhaled epoprostenol, inhaled nitric oxide) may be considered. However, these agents may have limited utility for improving systemic oxygenation when the ventilator is set to an ultra-lung-protective strategy and there is minimal alveolar ventilation. (See "Acute respiratory distress syndrome: Fluid management, pharmacotherapy, and supportive care in adults", section on 'Inhaled pulmonary vasodilators'.)

Managing concomitant cardiac failure — We treat concomitant cardiogenic shock with the usual medical therapies for circulatory support (eg, inotropic support).

If cardiogenic shock is refractory to medical management, we have a low threshold to add an arterial reinfusion limb (ie, conversion to venovenoarterial ECMO (figure 4)) to provide circulatory support. Details on venoarterial ECMO are provided separately. (See "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)".)

WEANING — Once the underlying disease process has sufficiently improved and it is assessed that the patient is adequately supported with conventional mechanical ventilation, we wean V-V ECMO. Criteria include improved lung compliance (eg, tidal volume 6 mL/kg predicted body weight [PBW] with plateau airway pressure ≤30 cm H₂O, in the setting of improving radiograph appearance, and adequate oxygenation at low fraction of inspired oxygen [FiO₂; eg, 0.3 to 0.5] without major changes in cardiac output).

In most patients, ECMO weaning involves transitioning to ongoing invasive mechanical ventilation. This may involve increases in the FiO₂, tidal volume, or respiratory rate to maintain acceptable oxygenation and ventilation in parallel with decreasing support from the extracorporeal device. Less commonly, patients may have been endotracheally extubated while receiving ECMO support and may be transitioned to spontaneous breathing with or without ventilatory assistance.

Our protocol is the following:

●Once sweep has been successfully decreased to a low level (eg, 1 L/minute), typically in conjunction with liberalized ventilator settings, we turn off the sweep gas flow (an "off-sweep challenge") for 30 minutes and obtain a systemic arterial blood gas (ABG). If the patient meets select criteria, we decannulate the patient.

●Reasonable criteria indicating readiness to decannulate from ECMO while off sweep gas flow includes the following:

•Patients receiving invasive mechanical ventilation:

-Oxygenation: FiO₂ ≤0.6 and positive end-expiratory pressure (PEEP) ≤10 cm H₂O while maintaining partial pressure of arterial oxygen (PaO₂) ≥60 mmHg

-Ventilation: pH ≥7.35 while being able to maintain a tidal volume ≤6 mL per kg of PBW with plateau airway pressure ≤30 cm H₂O; respiratory rate ≤28 without evidence of excessive work of breathing

•Patients endotracheally extubated:

-PaO₂ ≥80 mmHg with a clinically acceptable dose of supplemental oxygen (eg, <6 liters low-flow oxygen), pH ≥7.35, respiratory rate ≤28 without excessive work of breathing

●For patients who do not meet these criteria, ECMO is resumed.

For patients who have a borderline outcome from the "off-sweep challenge" (eg, patients requiring the upper limit of acceptable criteria to achieve the desired pH and PaO₂), there are several options. These include:

●An increase in the time allotted for the "off-sweep challenge" (eg, up to 24 hours).

●Incremental decreases in sweep gas flow or fraction of delivered oxygen (FDO₂) via the blender. Such changes may be on the order of 20 percent at a time, with peripheral ABG assessments performed approximately 30 minutes after any given decrease in support (or peripheral oxygen saturation checked in the case of weaning FDO₂).

Anticoagulation should be discontinued for one to two hours prior to decannulation. In order to preserve blood volume, blood from the ECMO circuit may be reinfused from the circuit to the patient at the time of decannulation, with consideration of concomitant diuresis if there is concern for development of volume overload [48]. After decannulation, clinicians either suture the insertion site or just hold pressure, which is at the discretion of the clinician or based on center policy.

Extubation and weaning from mechanical ventilation are similar to that in other critically ill patients and are discussed separately. (See "Weaning from mechanical ventilation: Readiness testing" and "Initial weaning strategy in mechanically ventilated adults" and "Extubation management in the adult intensive care unit".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Acute respiratory failure and acute respiratory distress syndrome in adults".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Extracorporeal membrane oxygenation (ECMO) (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Terminology – Venovenous (V-V) extracorporeal membrane oxygenation (ECMO) is a form of extracorporeal life support. V-V ECMO is mostly used for temporary support of patients with acute respiratory failure. The term V-V refers to the configuration of the circuit, which is comprised of a drainage cannula that removes blood from a large central vein (V-V). This blood is oxygenated extracorporeally and reinfused into the same or a different vein (V-V) (table 1). (See 'Terminology' above and "Extracorporeal life support in adults in the intensive care unit: Overview", section on 'Terminology'.)

●Patient selection – Candidates for V-V ECMO include patients who have severe, acute, and potentially reversible hypoxemic or mixed hypoxemic and hypercapnic respiratory failure, despite optimization of conventional management (eg, lung-protective ventilation [≤6 mL/kg predicted body weight and plateau airway pressure 30 cm H₂O], prone positioning, neuromuscular blockade). Examples include patients with unacceptably high airway pressures (eg, plateau pressure >32 cm H₂O) or patients who demonstrate intolerance of lung-protective ventilation, such as refractory hypoxemia or severe acidemia. (See 'Clinical applications' above.)

•The most common indication is acute respiratory distress syndrome (ARDS) (algorithm 1). For patients with severe ARDS refractory to conventional therapies, we suggest transfer to a center with an ECMO center (Grade 2C). If the expertise is available locally, we suggest ECMO, provided criteria are met (Grade 2C). (See 'Severe ARDS' above.):

These criteria include the following:

-Partial pressure of arterial oxygen to fraction of inspired oxygen [PaO₂:FiO₂] <50 mmHg for >3 hours

-PaO₂:FiO₂ <80 mmHg for >6 hours

-pH <7.25 and partial pressure of arterial carbon dioxide ≥60 mmHg for >6 hours (with respiratory rate at 35 breaths per minute and mechanical ventilation settings adjusted to keep plateau airway pressure ≤32 cm H₂O)

Other indications include needing ECMO support as a bridge to lung transplantation, primary graft dysfunction after lung transplantation, reperfusion pulmonary edema and cardiopulmonary failure following pulmonary thromboendarterectomy, and diffuse alveolar hemorrhage. (See 'Others' above and 'Severe lung disease with ECMO as a bridge to lung transplantation' above.)

•There are few absolute contraindications except for a pre-existing condition that is either incompatible with recovery or has no potential for a long-term life-saving therapy (eg, lung transplantation). Absolute and relative contraindications are listed in the table (table 3). (See 'Assess contraindications and technical challenges' above.)

●Catheter insertion – We commonly use a two-site cannulation approach under ultrasound guidance (figure 1). We select cannula sizes based on the patient's anticipated extracorporeal blood flow needs (eg, 29 French drainage cannula and 22 French reinfusion cannula for ARDS). Placement is performed by an expert in cannulation (eg, surgeon or cardiac interventionalist) under image-guidance, and a chest radiograph is performed after insertion to check for appropriate placement. (See 'Catheter insertion' above.)

●Initial settings and titration – Following cannulation, the patient is immediately connected to the ECMO circuit. Although variable, typical initial settings are achieved incrementally over the first few minutes after initiation (see 'Initial settings and titration' above):

•Pump speed increased to achieve an extracorporeal blood flow of at least 3 L/minute (range 3 to 7 L/minute)

•Sweep gas flow: typical range 1 to 10 L/minute, with initial settings usually at the lower end of that range

•Fraction of delivered oxygen of 1.0

Within an hour of cannulation, we generally obtain routine laboratories, systemic arterial blood gas (ABG; eg, radial ABG), coagulation studies, and pre- and post-membrane lung blood gases.

●Supportive therapies

•Mechanical ventilation settings – In parallel with titration of ECMO settings, we begin to adjust mechanical ventilation settings. For most patients, we suggest an "ultra-lung-protective" ventilator strategy (table 5) rather than the typical low-volume, low-pressure strategy (Grade 2C). This approach is based upon the ventilatory strategy used in one of the major ECMO ARDS trials that we also apply to non-ARDS patients. (See 'Invasive mechanical ventilation' above.)

•Anticoagulation – Patients receiving V-V ECMO require anticoagulation due to the high risk of systemic and circuit thrombosis. The agent used, target level of anticoagulation, and monitoring vary. Anticoagulation is typically performed using unfractionated heparin (UFH). We typically administer an initial bolus of UFH (eg, 5000 units) just prior to cannula insertion and start a continuous intravenous UFH infusion postcannulation. We suggest using a low-level anticoagulation strategy targeting an activated partial thromboplastin time of approximately 40 to 60 seconds (Grade 2C). (See 'Anticoagulation' above.)

●Monitoring – To ensure a properly functioning circuit and detect complications early, we examine patients daily and periodically throughout the day. We obtain daily complete blood count and chemistries, systemic ABG, coagulation studies, and chest radiography. If hemolysis is suspected, we obtain plasma-free hemoglobin, urinalysis, and peripheral smear. ECMO circuits are continuously monitored to ensure all components are functioning properly using pre- and post-membrane line pressures, drainage line pressure, cannula position, and signs of infection or thrombus. (See 'Daily assessment and circuit monitoring' above.)

●Complications – Common complications are bleeding, thrombocytopenia, thrombosis, hemolysis, vascular injury, intracranial bleeding or stroke, and infections. These and other complications are listed in the table (table 6). (See 'Complications' above.)

●Weaning – We consider weaning when patients have improved lung compliance (tidal volume 6 mL/kg predicted body weight with plateau airway pressure ≤30 cm H₂O), improving radiograph appearance, and adequate oxygenation at low fraction of inspired oxygen. Weaning in the majority of patients involves turning off the sweep gas flow (an "off-sweep challenge") for 30 minutes and obtaining a systemic ABG. If the patient meets select criteria, we decannulate the patient (See 'Weaning' above.)