The use of extracorporeal CO2 removal in acute respiratory failure

Background Chronic obstructive pulmonary disease (COPD) exacerbation and protective mechanical ventilation of acute respiratory distress syndrome (ARDS) patients induce hypercapnic respiratory acidosis. Main text Extracorporeal carbon dioxide removal (ECCO2R) aims to eliminate blood CO2 to fight against the adverse effects of hypercapnia and related acidosis. Hypercapnia has deleterious extrapulmonary consequences, particularly for the brain. In addition, in the lung, hypercapnia leads to: lower pH, pulmonary vasoconstriction, increases in right ventricular afterload, acute cor pulmonale. Moreover, hypercapnic acidosis may further damage the lungs by increasing both nitric oxide production and inflammation and altering alveolar epithelial cells. During an exacerbation of COPD, relieving the native lungs of at least a portion of the CO2 could potentially reduce the patient's respiratory work, Instead of mechanically increasing alveolar ventilation with MV in an already hyperinflated lung to increase CO2 removal, the use of ECCO2R may allow a decrease in respiratory volume and respiratory rate, resulting in improvement of lung mechanic. Thus, the use of ECCO2R may prevent noninvasive ventilation failure and allow intubated patients to be weaned off mechanical ventilation. In ARDS patients, ECCO2R may be used to promote an ultraprotective ventilation in allowing to lower tidal volume, plateau (Pplat) and driving pressures, parameters that have identified as a major risk factors for mortality. However, although ECCO2R appears to be effective in improving gas exchange and possibly in reducing the rate of endotracheal intubation and allowing more protective ventilation, its use may have pulmonary and hemodynamic consequences and may be associated with complications. Conclusion In selected patients, ECCO2R may be a promising adjunctive therapeutic strategy for the management of patients with severe COPD exacerbation and for the establishment of protective or ultraprotective ventilation in patients with ARDS without prognosis-threatening hypoxemia.


Background
Extracorporeal carbon dioxide removal (ECCO 2 R) is a technique whose objective is the decarboxylation of blood and thus to correct hypercapnia and respiratory acidosis [1,2]. ECCO 2 R is similar to extracorporeal membrane oxygenation (ECMO) but uses lower blood flow, usually less than 1500 mL/min. Therefore, this technique has little or no impact on blood oxygenation. Initially, ECCO 2 R was developed in the treatment of patients with acute respiratory distress syndrome (ARDS) [3], but because of the progressive improvement of this technique and its use in hospitals, ECCO 2 R could be proposed as a therapeutic option in cases of hypercapnic respiratory insufficiency, either during acute and severe decompensation of chronic obstructive pulmonary disease (COPD) [4] or in ARDS to achieve less invasive mechanical ventilation (IMV) [5]. In this review of the literature, we will discuss the current knowledge on the pathophysiology related to hypercapnic respiratory failure, the principles of the ECCO 2 R technique, and its place in the treatment of ARDS and acute and severe decompensations of COPD.

ECCO 2 R: from applied physiology to clinical studies
Pathophysiological rationale of the use of ECCO 2

R in COPD exacerbations
The amount of CO 2 in the blood is higher than that of oxygen. CO 2 is mainly present in blood as bicarbonates and to a lesser extent in dissolved form, whereas O 2 is mainly linked to hemoglobin. Small variations in the partial pressure of CO 2 (PaCO 2 ) cause significant variations in the level of CO 2 in the blood, unlike the relationship between the O 2 partial pressure and O 2 blood content. Therefore, extracorporeal CO 2 removal can be realized with lower blood flow rates than requires extracorporeal oxygenation but with enough fresh gas flow sweeping the exchange membrane [6]. ECCO 2 R aims to eliminate blood CO 2 to fight against potential adverse effects of hypercapnia and related acidosis. Hypercapnia has deleterious extrapulmonary consequences, particularly on the brain, by increasing cerebral blood flow and therefore intracranial pressure [7]. In addition, in the lungs, hypercapnia leads to pulmonary vasoconstriction, increases right ventricular afterload, and decreases myocardial contractility with consequent right heart failure [8]. Moreover, hypercapnic acidosis may further damage the lungs by increasing both nitric oxide production and inflammation and altering alveolar epithelial cells [9]. Finally, because of its immunosuppressive properties, hypercapnic acidosis may exacerbate lung damage by exacerbating pulmonary bacterial infections [9].
During exacerbations of COPD, the volume of CO 2 removed by the lungs is reduced due to worsening dynamic overdistension and the gap between ventilation and perfusion [10], accompanied by severe hypercapnia. In addition, in patients with COPD exacerbation, CO 2 production is estimated to be 23% higher than the normal value of 200 to 250 mL/min due to increased respiratory muscle work and metabolism [10].
Therefore, during an exacerbation of COPD, relieving the native lungs of at least a portion of the CO 2 could potentially improve the acid-base balance and reduce the patient's respiratory work, resulting in a reduced respiratory rate and alveolar ventilation [11]. Instead of mechanically increasing alveolar ventilation with IMV in an already hyperinflated lung to increase CO 2 removal, the use of ECCO 2 R may allow a decrease in respiratory volume and respiratory rate, resulting in longer expiratory time that is better adapted to the high expiratory time constant of the respiratory system. Through these physiological mechanisms, ECCO 2 R can neutralize the vicious cycle of dynamic hyperinflation and its harmful respiratory and cardiovascular consequences. Beneficial effects derived from respiratory mechanics, ventilatory muscle efficiency, respiration, and cardiovascular function can improve gas exchange and relieve dyspnea, potentially preventing the failure of NIV or facilitating weaning from IMV [10][11][12]. The pathophysiological rationale for the use of ECCO 2 R in COPD exacerbation is presented in Fig. 1 (Pathophysiology of respiratory acidosis is presented in Additional file 1 and pathophysiology of COPD is presented in Additional files 1 and 2 (Figure S1)).

Pathophysiological rationale of the use of ECCO 2 R in ARDS
In recent decades, very important progress has been made in the understanding of the pathophysiology of ARDS. The recognition of ventilatory-induced lung injury (VILI) has led to drastic changes in the ventilatory management of these patients [13,14]. The historical trial conducted by the ARDSNet group demonstrated that the ventilation of ARDS patients with a low tidal volume (VT) of 6 mL/kg (vs. 12 mL/kg) significantly reduced mortality [15]. However, recent results have shown that pulmonary hyperinflation still occurs in approximately 30% of ARDS patients despite this so-called "protective" ventilation [16]. This analysis suggests a beneficial effect of VT reduction, even in patients already at a plateau pressure (Pplat) < 30 cm H 2 O [17]. The decrease in the VT and Pplat will also decrease the driving pressure, which has recently been identified as a major risk factor for mortality in ARDS patients [18]. A reduction in VT to less than 6 mL/kg to reach a low Pplat level may induce severe hypercapnia that may increase intracranial pressure, causes pulmonary hypertension, decreases myocardial contractility, reduce renal blood flow, and releases endogenous catecholamines [19,20]. In a recent multicenter study on 35 ARDS patients with PaO 2 / FiO 2 ≤ 150 mmHg, Richard et al. reduced VT to 4 mL/ kg and further adjusted respiratory rate (RR) to keep pH ≥ 7.20. RR was augmented up to 40 breaths/min. On day 2, VT decreased from 6.0 [5.9-6.1] to 4.1 [4.0-4.7] ml/kg leading to a significant decrease in driving pressure from 12 [9][10][11][12][13][14][15] to 8 [6][7][8][9][10][11] cmH 2 O. They concluded that ultra-low tidal volume ventilation may be applied in approximately 2/3 of moderately severe-to-severe ARDS patients while 2 patients (6%) developed acute cor pulmonale and 11 patients (32%) developed transient severe acidosis with pH < 7.15. A 4 cmH 2 O median reduction in driving pressure has been reached, at the price of transient episodes of severe acidosis [21]. This strategy is therefore not feasible for most ARDS patients with conventional IMV [22]. Therefore, ECCO 2 R could be used to achieve a VT < 6 mL/kg, thus lowering the Pplat, driving pressure and mechanical power [23][24][25][26][27] while maintaining PaCO 2 and pH in physiological standards.

Technical principles
Catheters or cannulas are needed to implement this technique. There are two categories of ECCO 2 R. The first category is the so-called arteriovenous technique, where the removal of CO 2 is possible without a pump. A femoro-femoral approach is used. This technique requires arterial and venous cannulation with 15 French cannulas. The blood flow inside the device depends exclusively on the cardiac output of the patient, which explains the great variability of the ability to oxygenate the patient. However, with a membrane surface of 1.3 m 2 ,  (Fig. 2). The venovenous technique uses low or very low blood flow. Currently, it is the venovenous technique that is conventionally used for ECCO 2 R. The pumps used are rollers, centrifugal or diagonal, electric or electromagnetic. Figure 3 shows a schematic representation of different ECCO 2 R systems. The gas exchange membrane is a device with a complex geometry based on hollow fibers. The material used is poly-4-methyl-1-pentene (PMP). The exchange surfaces vary in size from 0.32 to 0.65 m 2 for venovenous systems and 1.3 m 2 for arteriovenous systems. Circuits such as membranes are coated with heparin for better biocompatibility, better gas exchange and less capillary leakage. The extraction of carbon dioxide is done through the sweeping of the membrane by a fresh gas (O 2 or medical air) devoid of CO 2 . Current systems used to remove CO 2 are venovenous and use double-lumen venous catheters/ cannulas. The venous approach is classically achieved through the right internal jugular or femoral vein, and puncture of the vessel is performed under ultrasound guidance. The placement of the guidewire and the cannula requires control by transesophageal or subxiphoid transthoracic echocardiography (Fig. 4). Anticoagulation therapy (anti-Xa activity between 0.3 and 0.6 IU/L) is mandatory to avoid thrombosis in the circuit. Thus, any patient with a contraindication to anticoagulation therapy cannot benefit from ECCO 2 R. There are different types of machines on the market. The devices adapted from the VV-ECMO technique are very effective for CO 2 removal but require the insertion of cannulas between 18 and 19 French. The blood flow generated is between 500 and 1500 mL/min. The newest ECCO 2 R devices are relatively simple to use because they require the insertion of a smaller double-lumen cannula (up to [13][14][15] and work with very low blood flow rates (between 0.2 to 0.5 L/min). However, their CO 2 removal performance remains limited [11]. The characteristics of the different ECCO 2 R systems available on the market are summarized in Table 1.

Use of ECCO 2 R in severe acute exacerbations of COPD
Noninvasive ventilation remains the gold standard for the treatment of acute hypercapnic respiratory failure [28], but in approximately 20 to 30% of cases, this technique may not be sufficient, and patients need to be intubated and mechanically ventilated. The mortality of patients requiring the use of IMV is higher than those receiving NIV alone. Thus, the combination of ECCO 2 R therapy with NIV could be a therapeutic option to reduce the failure of NIV and prevent the use of intubation and IMV. In fact, the use of ECCO 2 R in patients with hypercapnic respiratory failure may improve the efficacy of NIV in terms, that ECCO 2 R, decreases respiratory rate, and reduces dynamic hyperinflation and intrinsic PEEP. In addition, by avoiding the use of IMV and therefore endotracheal intubation, it is also possible to limit the adverse effects related to analgo-sedation, which include hemodynamic instability, difficult and prolonged respiratory weaning, and a significant number of neurological disorders related to prolonged sedation. The absence of analgosedation also allows patients to drink, eat, communicate with relatives, and perform active physiotherapy. In addition, it has recently been demonstrated that ECCO 2 R, by decreasing the respiratory rate, can reduce the work of breathing and decrease the CO 2 production of the respiratory muscles. Therefore, this contributes to the decrease in PaCO 2 [29]. As a result, this may facilitate weaning from IMV and promote earlier extubation.

Use of ECCO 2 R decreases the use of IMV in patients with COPD exacerbation
Kluge et al. [30] investigated the feasibility of a pumpless extracorporeal assist (PECLA) system in 21 patients with COPD who did not respond to NIV. The use of PECLA system was associated with decreased PaCO 2 levels and improved pH after 24 h and avoided the use of intubation and IMV in 90% of treated patients. Retrospective analysis with a control group showed no significant difference in mortality at 28 days (19% with ECCO 2 R and 24% without ECCO 2 R) or at 6 months (both groups at 33%) or in the median duration of ICU or hospital length of stay (15 vs 30 days and 23 vs 42 days, respectively). In the study conducted by Burki et al. [4], 20 hypercapnic patients with COPD were treated with ECCO 2 R using a 15.5-Fr dual-lumen cannula, allowing an average blood flow of 430 mL/min. The authors reported improvement in both hypercapnia and respiratory acidosis, and IMV was avoided in the nine patients treated with NIV. More recently, Del Sorbo et al. [31] reported that ECCO 2 R with a 14-Fr dual-lumen catheter and blood flow rates of 177 to 333 mL/min not only improved respiratory acidosis, but also reduced the need for intubation in 25 patients with COPD who have a high risk of NIV failure. Compared with the control group, intubation risk and hospital mortality were significantly lower in the ECCO 2 R group. These results were challenged in a recent study by Braune et al. [32], which showed that IMV was avoided in 56% of patients treated with ECCO 2 R but was associated with a higher incidence of complications. However, several differences were found between these two studies, including the inclusion of patients with contraindications for NIV and the unexpectedly high incidence of hypoxemic patients [33]. In another study, Morelli et al. [34] confirmed the efficacy of ECCO 2 R (with a flow rate of 250 to 450 mL/min via a 13-Fr dual-lumen catheter) to reduce PaCO 2 in a case series of 30 patients with acute hypercapnic respiratory failure due to COPD exacerbation who refused endotracheal intubation after NIV failure. The duration of ECCO 2 R treatment was 2 to 16 days, and it was possible to avoid endotracheal intubation in 27 patients. Finally, in a round table, 14 European experts' views were collated to better understand how ECCO 2 R therapy is used, how patients are selected and managed. In COPD patients with acute exacerbation, a consensus was found that, in patients at risk of NIV failure, no decrease in PaCO 2 and no decrease in respiratory rate were principal criteria for starting with ECCO 2 R therapy. Main treatment targets in COPD patients were patient well-being, pH (> 7.30-7.35), respiratory rate (< 20-25 breaths/min), decrease of PaCO 2 (by 10-20%), weaning from NIV, decrease in HCO 3 − and maintaining hemodynamic stability [35].

Use of ECCO 2 R to help weaning from IMV
In the case series of Elliot et al. of two patients with severe acute asthma [36], the addition of pumpless ECCO 2 R to IMV corrected hypercapnia and related acidosis and reduced other support measures, including hemodynamics, and allowed weaning from IMV. In the study by Burki et al. [4], in the subgroup of 11 mechanically ventilated patients, ECCO 2 R allowed weaning from IMV in only 3 patients. Nevertheless, even if they were not completely weaned, in three other patients, ventilatory assistance could be reduced. Using a double-lumen cannula (20-23 Fr) with a blood flow of 1 to 1.7 L/min, Abrams et al. [37] successfully weaned and extubated five COPD patients with acute respiratory acidosis after only 24 h of IMV. All of these patients survived until discharge from the hospital. Similarly, using a pediatric VV-ECMO system (with blood flow rates of 0.9 L/min through a 19-Fr double-lumen cannula placed in the right internal jugular vein) in two adult patients with COPD exacerbation, Roncon-Albuquerque Jr [38] reported early extubation after 72 h and patient mobilization on day 6. A retrospective analysis of data from 12 patients with hypercapnic respiratory failure confirms the efficacy of ECCO 2 R with median blood flow rates of 1.2 to 1.4 L/min in the correction of hypercapnia and in the reduction of ventilation pressures and minute ventilation. Of the patients studied, six patients with mainly hypercapnic pulmonary insufficiency due to COPD or fibrosis were promptly weaned off of IMV and survived until discharge from the hospital. It should be noted that five patients were awake and spontaneously breathing during ECCO 2 R therapy [39].
Taken together, these results support the notion that ECCO 2 R may be useful for the avoidance of intubation during NIV and for the facilitation of weaning from IMV. Nevertheless, the observational nature of the available data makes it difficult to understand the efficacy and safety of such strategies in these patients.
The relevant clinical studies on ECCO 2 R in COPD are summarized in Table 2.

Use of ECCO 2 R in acute respiratory distress syndrome (ARDS)
The latest feasibility and safety pilot study of 20 patients with moderate and/or severe ARDS, in whom ECCO 2 R was performed with a new standalone platform (without concomitant extrarenal treatment), Prismalung ® (Gambro-Baxter), integrated on the Prismaflex ® platform (Gambro-Baxter), showed a reduction in the tidal volume from 6 to 4 mL/kg of the predicted body weight and in Pplat below 25 cmH 2 O, thus achieving ultraprotective ventilation. However, the results show that despite maximal ECCO 2 R treatment (sweep gas flow at 10 ± 0.3 L/min and blood flow at 421 ± 40 mL/min, corresponding to the maximum that this platform can generate), patients ventilated at 4 mL/kg of their predicted body weight become acidotic (pH decreased from 7.39 ± 0.1 to 7.32 ± 0.10 and PaCO 2 increased from 43 ± 8 mmHg to 53 ± 9 mmHg) [5].
A larger prospective multicenter international phase II study aimed to assess the feasibility and safety of extracorporeal carbon dioxide removal (ECCO 2 R) to facilitate ultraprotective ventilation (V T 4 mL/kg and Pplat ≤ 25 cmH 2 O) in patients with moderate ARDS. The primary endpoint was the proportion of patients achieving ultraprotective ventilation with PaCO 2 not increasing more than 20% from baseline and arterial pH > 7.30. Both lower CO 2 extraction and higher CO 2 extraction devices (membrane lung cross-sectional area 0.59 vs. 1.30 m 2 ;     [35]. Finally, using data from the SUPERNOVA trial (95 patients with early moderate ARDS), Goligher et al. assessed the independent effects of alveolar dead space fraction (ADF), respiratory system compliance (Crs), hypoxemia (PaO 2 /FiO 2 ), and device performance (higher vs lower CO 2 extraction) on the magnitude of reduction in V t , driving pressure and mechanical power permitted by ECCO 2 R were assessed. The authors demonstrated that patients with higher ADF or lower Crs and patients treated with higher CO 2 extraction are most likely to benefit from ECCO 2 R [41]. The combination of continuous renal replacement therapy (CRRT) and ECCO 2 R with very low blood flow is a promising concept. The hypothesis in a study by Moerer et al. is that this combined system can remove enough CO 2 to facilitate protective ventilation in mechanically ventilated patients. In 11 ventilated patients with acute renal failure placed under CRRT, a very-low-flow ECCO 2 R (300 mL/min) was added to the circuit. During 6 h of combined therapy, the elimination of CO 2 and its effect on the possibility of achieving protective ventilation were evaluated. The ventilation settings were maintained in assisted mode or in controlled pressure mode, allowing spontaneous breathing. With very-low-flow ECCO 2 R, a significant decrease in minute ventilation, tidal volume and paCO 2 was possible after 1-3 h but not after 6 h of treatment. On the other hand, no significant reduction in the driving pressure was observed during the combined treatment. The CO 2 removal was 20.73 mL CO 2 /min. Therefore, the very low blood flow in ECCO 2 R associated with CRRT treatment is not enough to significantly reduce respiratory work. The absolute cause could be the absolute amount of CO 2 removed by approximately 10% of CO 2 production in the resting adult. Therefore, the effectiveness of ECCO 2 R with very low blood flow in allowing protective ventilation is very limited [42]. Moreover, in another recent study including 20 hypercapnic ARDS patients requiring CRRT who were treated with a system combining very-low-flow ECCO 2 R (membrane lung 0.32 m 2 ) and renal replacement therapy, the pH increased from 7.18 ± 0.09 to 7.22 ± 0.08 (p < 0.05). There was a significant reduction in ventilation requirements with a decrease in tidal volume from 6.2 ± 0.9 to 5.4 ± 1.1 mL/kg PBW (p < 0.05), associated to a reduced pulmonary stress and strain [43]. Even if these results were statistically significant, we can question their clinical relevance. The relevant clinical studies on ECCO 2 R in ARDS are summarized in Table 3.

Role of ECCO 2 R while awaiting lung transplantation
It is well known that patients who develop acute gas exchange impairment requiring IMV while awaiting lung transplantation are more likely to die than patients who do not require IMV [44]. The reason for using ECCO 2 R in such patients is the possibility of the avoidance of endotracheal intubation and IMV, thus limiting their adverse effects (i.e., ventilator-associated pneumonia) that may preclude transplantation. In addition, by using ECCO 2 R, it is possible to avoid analgo-sedation, which allows the patient to maintain the tone of the respiratory muscles and to continue to perform active physiotherapy. Despite this pathophysiological rationale, studies regarding the use of ECCO 2 R in this subgroup of hypercapnic patients are still rare. Schellongowski et al. [45] performed a retrospective study of 20 patients with bronchiolitis obliterans, cystic fibrosis and idiopathic pulmonary fibrosis with indication for lung transplantation (n = 13) or retransplantation (n = 7). The use of venovenous ECCO 2 R and pumpless arteriovenous ECCO 2 R was associated with an improvement in hypercapnia and acidosis during the first 12 h of treatment. After a transition period of 4 to 11 days, 19 patients (95%) were successfully transplanted. Survival at the hospital was 75%. A very recent retrospective study confirmed that patients treated with ECCO 2 R before lung re-transplantation tended to have better survival [46]. In light of these findings, ECCO 2 R may even be useful in thoracic surgical procedures other than lung transplantation [47]. Nevertheless, given the complexity and the difficult clinical conditions of these patients awaiting lung transplantation, the use of ECCO 2 R in these patients should be performed only in experienced centers.

ECCO 2 R-related complications and technical limitations
The use of ECCO 2 R may have pulmonary and hemodynamic consequences and may be associated with complications. Adverse events include events related to the patient, the circuit and mechanical events (Table 4). In four studies of ARDS patients, the use of ECCO 2 R was associated with hypoxemia and the need for an increase in FiO 2 . The present fact could be explained by lung derecruitment related to decrease in ventilation (favoring atelectasis). Moreover, PaO 2 /FiO 2 worsening in ECCO 2 R may at least in part, reflect a modification of the alveolar gas content due to ECCO 2 R (modification of the respiratory quotient) [48]. To correct this phenomenon, IMV was implemented in spontaneously breathing patients [49] with both the use of high levels of PEEP and prone position to maintain functional residual capacity [24,[49][50][51]. In case of refractory hypoxemia a switch to VV-ECMO [52] was performed.
The major adverse effects may be caused by venous and/or arterial cannulation, with increased risk depending on the choice of vascular access and the type and size of cannulas. Transient ischemia of the lower limb, "false" aneurysm of the femoral artery [50] and fatal perforation following retroperitoneal bleeding have been described [4,33].
Anticoagulation protocols with heparin are necessary to maintain the efficacy and performance of ECCO 2 R [53]. Thus, hemorrhagic events may be considered the most common complication and are associated with a higher number of blood transfusions during ECCO 2 R therapy [4,30,33,49,50,52].
Transient thrombocytopenia, probably related to the use of heparin, has also been noted [4,33,51]. However, thrombocytopenia and decreased coagulation factors, certainly due to an activation of coagulation and fibrinolysis as well as an inflammatory response mediated by the complement system [54] may also be the result of interactions between blood components and the circuit. Future research should focus on improvements in anticoagulation protocols and the development of practical guidelines [55].
Despite anticoagulation protocols, clot formation in the circuits often occurs reducing the clearance of CO 2 in the membrane and resulting in a rapid increase in PaCO 2 . The occurrence of membrane thrombosis should be considered a life-threatening event and necessitates rapid circuit changes, changes in ventilator parameters, and endotracheal intubation in the case of NIV [33,51,52]. Moreover, it seems that the reduction in blood flow through the membrane may be linked to an increase in the risk of thrombosis of the system. In the study of Schmidt et al. including 20 patients with mild or moderate ARDS, VT was gradually lowered from 6 to 5, 4.5, and 4 ml/kg. When arterial PaCO 2 increased by > 20% from its initial value, a very-low-flow standalone ECCO 2 R was initiated to reduce respiratory acidosis. The authors showed that despite a heparin-infusion protocol that also included a bolus at catheter insertion, 50% of the treated patients experienced membrane clotting before the end of the experimental protocol [5]. In a retrospective study carried out by our group on 3 patients with severe COPD also assisted by a very-low-flow ECCO 2 R, thrombosis of the circuit occurred in 2 patients. In contrast, in our study, the 6 patients assisted by a higher blow flow ECCO 2 R did not experience circuit thrombosis [56]. It therefore appears that the blood flow passing throughout the circuit has a role in the occurrence of circuit thrombosis.
The displacement or twisting of the catheter/cannula may cause pump malfunction and promote thrombosis of the membrane. Finally, episodes of intravascular hemolysis have been reported in two case series, including one requiring a transfusion [51,52].
Finally, CO 2 extraction capacity differed between the devices available on the market. While re-analyzing the results of the SUPERNOVA trial according to the ECCO 2 R devices used (lower blood flow (area of membrane length 0.59 m2; blood flow 300-500 mL/min) vs higher blood flow (membrane area 1.30 m 2 ; blood flow between 800 and 1000 mL/min), Combes et al. showed that reduction of V T to 4 mL/kg was achieved in 55% and 64% of patients with the lower extraction versus 90% and 92% of patients with higher extraction devices at 8 and 24 h from baseline, respectively (p < 0.001) [57]. Moreover, ECCO 2 R-related hemolysis and bleeding were higher with lower than with higher extraction devices. In our retrospective study on COPD patients, we showed that when compared with a higher blood flow ECCO 2 R system, very low-flow device was not able to remove sufficient CO 2 , normalize pH or decrease respiratory rate [56].

New technologies and ongoing research on ECCO 2 R
ECCO 2 R devices remove CO 2 directly from the blood, facilitating ultraprotective ventilation or even offering an alternative to IMV. However, ECCO 2 R is not widely available, while dialysis is available in most intensive care units. Recent technological advances are focused on the development of minimally invasive devices that provide adequate CO 2 removal with increased safety and simple use. Previous attempts to perform ECCO 2 R with dialysis by removing CO 2 as bicarbonate have been affected by metabolic acidosis. Bicarbonate dialysis is possible, provided that the difference between the strong ions in the plasma is maintained. Using a mathematical model        to study the effects of bicarbonate removal on pH and CO 2 in plasma, in vitro experiments were performed to test CO 2 removal using three dialysates with different bicarbonate concentrations (0, 16 and 32 mmol/L). This model predicts a reduction in partial CO 2 pressure (PaCO 2 ) and an increase in pH with a progressive reduction in plasma bicarbonate, provided that the strong ion difference and the maintenance of plasma proteins are preserved. In these in vitro experiments, CO 2 removal with an adult-size filter was maximal with a dialysate not containing bicarbonate, equivalent to 94 mL/min (± 3.0) of CO 2 eliminated. Under the same conditions, the dialysate containing a conventional concentration of bicarbonates (32 mmol/L) eliminated only 5 mL/min (± 4, p < 0.001). As expected, the pH increased after the removal of the bicarbonate. These data show that dialysis with low-bicarbonate dialysates is feasible and results in a reduction in plasma PaCO 2 . When scaled to estimate equivalent CO 2 removal with an adult dialysis circuit, the amount eliminated competes with that of existing low-flow ECCO 2 R devices [58]. However, these methods may be impractical for clinical use due to acid-base disturbances, hemolysis, cardiac arrhythmias and micronutrient depletion despite several attempts to replace bicarbonate [59,60]. Finally, other techniques were evaluated, including the combination of ECCO 2 R and continuous renal replacement therapy, the acidification of blood with lactic acid, the addition of carbonic anhydrase to the membrane and electrodialysis [60][61][62]. ECCO 2 R technique based on infusion of metabolizable acids exploits bicarbonate for gas exchange. An innovative lung support technique, called respiratory electrodialysis has been developed, consisting in a combination of a hemofilter, a membrane lung, and an electrodialysis unit. By applying electrodialysis to hemodiafiltrate, the pH and the electrolyte concentration are selectively modulated in specific sections of the extracorporeal circuitry. Blood is regionally acidified, bicarbonate is exchanged with chloride, and the PaCO 2 is increased, leading to facilitated membrane lung CO 2 removal [61]. These strategies can enhance the physiological benefits of ECCO 2 R while reducing its risks. However, studies demonstrating safety and efficacy are necessary before putting these technological innovations into clinical practice. Several studies of ECCO 2 R are currently underway in patients with hypercapnic respiratory failure (Clinical-Trials.gov). Details of these studies are available in Additional file 2. These various ongoing clinical studies on the use of ECCO 2 R in COPD and ARDS are summarized in Additional file 2: Tables S1, S2, respectively.