Comments
Abstract
Rationale: Weaning from venovenous extracorporeal membrane oxygenation (VV-ECMO) is based on oxygenation and not on carbon dioxide elimination.
Objectives: To predict readiness to wean from VV-ECMO.
Methods: In this multicenter study of mechanically ventilated adults with severe acute respiratory distress syndrome receiving VV-ECMO, we investigated a variable based on CO2 elimination. The study included a prospective interventional study of a physiological cohort (n = 26) and a retrospective clinical cohort (n = 638).
Measurements and Main Results: Weaning failure in the clinical and physiological cohorts were 37% and 42%, respectively. The main cause of failure in the physiological cohort was high inspiratory effort or respiratory rate. All patients exhaled similar amounts of CO2, but in patients who failed the weaning trial, e was higher to maintain the PaCO2 unchanged. The effort to eliminate one unit-volume of CO2, was double in patients who failed (68.9 [42.4–123] vs. 39 [20.1–57] cm H2O/[L/min]; P = 0.007), owing to the higher physiological Vd (68 [58.73] % vs. 54 [41.64] %; P = 0.012). End-tidal partial carbon dioxide pressure (PetCO2)/PaCO2 ratio was a clinical variable strongly associated with weaning outcome at baseline, with area under the receiver operating characteristic curve of 0.87 (95% confidence interval [CI], 0.71–1). Similarly, the PetCO2/PaCO2 ratio was associated with weaning outcome in the clinical cohort both before the weaning trial (odds ratio, 4.14; 95% CI, 1.32–12.2; P = 0.015) and at a sweep gas flow of zero (odds ratio, 13.1; 95% CI, 4–44.4; P < 0.001).
Conclusions: The primary reason for weaning failure from VV-ECMO is high effort to eliminate CO2. A higher PetCO2/PaCO2 ratio was associated with greater likelihood of weaning from VV-ECMO.
Weaning from venovenous extracorporeal membrane oxygenation is often based on the ability to maintain oxygenation during a progressive reduction and cessation of sweep gas flow. Despite the recognition that carbon dioxide elimination plays an important role in weaning failure, there is currently no guidance on the parameters able to identify patients who will successfully wean from extracorporeal membrane oxygenation based on exchange of carbon dioxide
This large multicenter physiological and clinical study shows that weaning failure is frequent and associated with the respiratory effort to control carbon dioxide exchange. The end-tidal partial carbon dioxide pressure/PaCO2 ratio was a strong predictor of weaning and can be readily used to increase the prediction of weaning success and increase safety.
The indications for initiating venovenous extracorporeal membrane oxygenation (VV-ECMO) in severe acute respiratory failure are based on defined oxygenation criteria that have remained substantially unchanged for the past four decades, dating from the first randomized trial by Zapol and colleagues (1) to the ECMO to Rescue Lung Injury in Severe ARDS trial by Combes and colleagues in 2018 (2). In contrast, the criteria and method of weaning from VV-ECMO are not well defined and may vary greatly regarding important aspects. For example: 1) the timing of weaning onset, considered to be the first reduction of sweep gas flow rate or extracorporeal blood flow; 2) the required observation time; and 3) the monitoring and assessment of key variables, such as the inspiratory effort, during the weaning procedure. In Table E1 in the online supplement, we summarize previously suggested protocols for weaning from VV-ECMO (2–15). The first requirement to initiate weaning is that the underlying clinical condition of the patient has convincingly improved—a clinical judgement that includes the trajectory of response to treatment. Using these criteria, weaning from VV-ECMO is pragmatically initiated following a trial-and-error approach based mainly on subjective clinical criteria rather than on objective data. The weaning trial is initiated and judged to be successful when the patient maintains adequate blood gases for an indefinite period without excessive respiratory effort or extracorporeal support. In the absence of hemodynamic or respiratory instability, weaning success or failure has generally been based on the following: 1) acceptable blood gas analysis, with FiO2 < 60% and ventilation either unaided or receiving modest pressure support; and 2) absence of excessive inspiratory effort, regardless of the magnitude of extracorporeal support, which usually is titrated to achieve oxygenation and PaCO2 values within normal ranges.
In the current study, we measured the variables most commonly used to assess weanability from ECMO, together with other variables not routinely measured, with the intent to: 1) define and quantify the physiological reasons for weaning failure; and 2) determine those variables highly predictive of weaning failure. To achieve these two aims we first studied a restricted physiological cohort to better understand the physiological reasons for failure to wean from extracorporeal support and to reveal possible predictive variables. We then examined a large clinical cohort to test the validity of those results.
This study included two cohorts: a prospective physiological cohort (including 26 patients) and a retrospective clinical cohort (including 638 patients). For the physiological cohort, we conducted a prospective interventional study in the ICU of the University Medical Center of Göttingen (Germany), from October 2017 until March 2020. We enrolled all patients >18 years old admitted to the ICU with the diagnosis of acute respiratory distress syndrome (ARDS) (16) and undergoing VV-ECMO. The study was approved by the institutional ethics board (Antragsnummer 09/06/17), and informed consent was obtained from the patient or from the legal representative. For the retrospective clinical cohort, we included the 638 patients who underwent VV-ECMO for severe respiratory failure at Guy’s and St. Thomas’ NHS Foundation Trust, London, United Kingdom, from January 2012 until November 2021. Figure 1 depicts the flow diagram of patients screened and included in the clinical cohort. The validation study had institutional approval (research governance reference number 7820), with the need for individual consent waived for the retrospective collection of data generated through routine care.
Figure 1. Flow diagram of patients included in clinical cohort. LOS = length of stay; VV-ECMO = venovenous extracorporeal membrane oxygenation.
[More] [Minimize]Extracorporeal support was set according to the Extracorporeal Life Support Organization guidelines (3, 17). In addition, in the physiological cohort, all patients were instrumented with an esophageal balloon catheter whose appropriate positioning was confirmed by the Baydur maneuver (18). The CO2 extracted (co2) by the natural and the membrane lungs was assessed using a prototype device (Dimar s.r.l.) that measures airflow and CO2 concentration from the ventilator and membrane lung.
Total body o2 was determined as the sum of two portions. The portion of o2 from the natural lung could not be measured directly, but it was estimated as the product of the arterial central venous O2 content difference multiplied by the cardiac output (computed as the ratio between the co2 of the natural lung and the delta central venous-arterial CO2 content; see online supplement). The portion of total co2 from the membrane lung was measured as the product of the extracorporeal blood flow and the arteriovenous O2 content difference of the membrane lung.
For the clinical cohort, we searched for all patients who underwent a VV-ECMO weaning trial over the last 10 years. We extracted data regarding demographics, ventilatory settings, VV-ECMO settings, gas exchange, and weaning outcome.
In the present study we refer only to the first weaning attempt in both cohorts, to provide the most homogenous data possible. In the physiological cohort, a trial of weaning from ECMO was attempted as soon as the patient, while receiving extracorporeal support, fulfilled the following weaning criteria: 1) assisted ventilation with pressure-support ventilation; 2) tidal swing of esophageal pressure (delta PES) ⩽ 15 cm H2O; 3) respiratory rate ⩽ 30 bpm; 4) arterial pH > 7.25; 5) PaCO2 ⩽ 60 mm Hg; 6) PaO2 ⩾ 70 mm Hg;with 7) FiO2natural lung ⩽ 60%. The weaning trial protocol consisted of a three-step reduction of the sweep gas flow from the original value (100%, baseline; 66%, step 1; 33%, step 2; 0%, step 3). Twenty minutes after each decrement of gas flow, a new measurement was performed, and the patient was eligible for further reduction of the gas flow if the weaning criteria were still fulfilled. If any one of weaning criteria was not met, the weaning trial was stopped and the ECMO settings were restored to the baseline settings (weaning failure). If the patient maintained all values of the weaning criteria at a sweep gas flow of 0 L/min (ECMO off), weaning was considered successful. We defined inspiratory effort as the inspiratory swings in esophageal pressure (delta PES) of ⩾15 cm H2O.
In the clinical cohort, the criteria for initiating weaning were the following: 1) resolution of underlying disease; 2) spontaneous breathing on pressure support ventilation (continuous positive airway pressure/pressure support); and 3) hemodynamic stability. During the VV-ECMO weaning test, the ventilator FiO2 was increased to 0.6 and sweep gas flow reduced in 33% steps every 10–15 minutes until sweep gas flow was at 0 L/min (ECMO off) with unchanging extracorporeal blood flow. The weaning test was interrupted and the patient returned to ECMO if any of the following failure criteria were met: 1) oxygen saturation as measured by pulse oximetry < 88% on FiO2 0.6; 2) airway occlusion pressure is the pressure generated at the airways during the first 100 msec of an inspiratory effort against an occluded airway (P0.1) > 10 cm H2O; 3) respiratory rate > 35 bpm; 4) any obvious signs of distress; or 5) increase in PaCO2 causing a decrease in pH < 7.35.
Data are expressed as mean ± SD or median (interquartile range), as appropriate. Comparison between two means or medians was performed with Student’s t test or Wilcoxon’s test. Linear regression was used to test linear relationships, and linear mixed model was used to account for repeated measures during the weaning steps. Receiver operating characteristics (ROC) curve analysis and logistic regression were used to test the effect of variables on outcome. Univariable logistic regression was used to estimate odds ratios (ORs) for individual parameters and measurements, with success or failure of the initial weaning trial as the dependent variable or outcome. ROC curves were constructed to assess the utility of each parameter in predicting the success of a weaning trial. Two-tailed P values < 0.05 were considered statistically significant.
All analysis were performed with R for Statistical Computing 4.0.
In the physiological cohort we obtained detailed data on the time course and outcome of weaning. This cohort comprised two groups: those who succeeded and those who failed the weaning attempt. All 26 patients adapted to the decrease in sweep gas flow by maintaining PaO2 and PaCO2 constant for the same total o2 and CO2 production.
As shown in Figure 2, upper panel, the PaO2 decreased slightly but significantly only when the sweep gas flow was zero in step 4 (sweep gas flow off), but it was not different between the two success or failure groups at any step. As shown in Figure 2, lower panel, the total o2 did not change significantly throughout the steps. However, the proportion of O2 exchanged through the natural lung progressively and significantly increased in the patients who failed the weaning process.
Figure 2. Upper panel: PaO2 as a function of the sweep gas flow. The PaO2 values were significantly different across the steps (step 4 vs. steps 1–3; P < 0.001) but similar between the groups who succeeded or failed weaning (P = 0.30). Lower panel: o2 as a function of the sweep gas flow. The total o2 (total height of the column) values were similar within the steps (P = 0.52) and between the groups (P = 0.25). The oxygen portion provided by the artificial lung was similar between groups (P = 0.20) but different between steps (P < 0.001). Values expressed as median ± interquartile range.
[More] [Minimize]As shown in Figure E1, the venous admixture was similar in the two groups and did not change significantly at any of the steps, including the final step in which the sweep gas flow was turned off.
As shown in the upper panel of Figure 3, PaCO2 changed throughout the decreases in sweep gas flow but remained similar between the successfully weaned and unweaned patients. Equally, the total co2 held constant across the four decremental steps of sweep gas flow, indicating that the decrease of the CO2 cleared by the membrane lung was fully compensated by an increase in the CO2 expired by the natural lung (Figure 3, lower panel).
Figure 3. Upper panel: PaCO2 as a function of the sweep gas flow. The PaCO2 values were significantly different across the steps (step 4 vs. steps 1–3; P < 0.001) but similar between the groups who succeeded or failed the weaning attempt (P = 0.4). Lower panel: co2 as a function of the sweep gas flow. The total co2 (total height of the column) values were similar across the steps (P = 0.52) and between the groups (P = 0.12). The CO2 eliminated by the natural lung and by the artificial lung at each step were different between the steps (P < 0.001 and P < 0.001, respectively) and groups (P = 0.039 and P = 0.027). Values expressed as median ± interquartile range.
[More] [Minimize]Arterial pressures and the central venous pressure were modestly but significantly lower throughout the weaning in failing patients compared with the successful ones (see online supplement for details).
In the physiological cohort, 42% of the patients failed weaning. In 70% of cases, this failure was due to excessive inspiratory effort (delta PES) and respiratory rate and in 30% of cases to hypoxemia (Table E2). The patients who failed weaning because of hypoxemia showed a higher median (interquartile range) baseline venous admixture fraction (Qva/Q) (0.28 [0.27–0.36] vs. 0.19 [0.18–0.3]; P = 0.040), whereas the patients failing for excessive inspiratory effort (delta PES) tended to have greater physiological Vd at baseline than the patients who succeeded (69.1% [58–82%] vs. 57% [45–66%]; P = 0.047). The higher physiological Vd measured in patients who failed to wean meant that to eliminate the same quantity of CO2 they were required to generate, throughout all weaning steps, inspiratory efforts (delta PES) that were almost double those of patients who passed the weaning test: 68.9 (42.4–123) cm H2O/(L/min) versus 39 (20.1–57) cm H2O/(L/min); P = 0.007 (Figure 4) (15, 19).
Figure 4. Left panel: Physiological Vd as a function of the membrane lung gas flow. The physiological Vd was different for patients who succeeded and for those who did not succeed the weaning attempt (P = 0.012). Physiological Vd fraction also differed across the different steps (P < 0.001). The decrease of physiological Vd during the steps was primarily due to a decrease of its anatomical component, as the ventilation through the natural lung increased during the steps in both groups (see online supplement for details, Figure E3). Right panel: Pleural pressure swing (i.e., the transpulmonary pressure swing per liter of exhaled CO2 [delta esophageal pressure [PES]/co2 natural lung]), as a function of the sweep gas flow. As shown, quite minor differences were found across the steps within each success or failure category. This relative stability indicates a near-constant relationship between delta of PES and co2 for each category, regardless of the total CO2 exhaled by the native lung. The magnitudes of that relationship, however, differed remarkably for patients who failed versus those who succeeded the weaning attempt. All data are expressed as median ± interquartile range.
[More] [Minimize]The most relevant baseline variables in physiological and clinical cohorts are shown in Table 1. The patients in the physiological cohort were more likely to be male and were treated with slightly higher positive end-expiratory pressure and FiO2, lower e, and higher sweep gas flow than were patients of the clinical cohort. The PaO2, PaCO2, and end-tidal partial carbon dioxide pressure (PetCO2)/PaCO2 ratio were similar in the two cohorts when the weaning began. The individual reasons for weaning failure were not available in the electronic database, but the cumulative failure rate was similar to the physiological cohort (37% in the clinical vs. 42% in the physiological; P = 0.5).
| Cohort | Physiological (n = 26) | Clinical (n = 638) | P Value |
|---|---|---|---|
| Age, yr | 57 (47–63) | 44 (34–53) | <0.001 |
| Male, n (%) | 21 (80) | 365 (57) | <0.001 |
| SOFA score | 12 (9–14) | 8 (7–12) | |
| Causes of ARDS, % | |||
| Pneumonia | 65 | 66.5 | — |
| Sepsis | 20 | 7.5 | — |
| Others | 15 | 26 | — |
| PaO2, mm Hg | 85.5 (73–103) | 88 (77–111) | 0.145 |
| PaO2/FiO2 | 220 (173–261) | 277 (223–340) | <0.001 |
| PaCO2, mm Hg | 44 (41–48) | 45 (42–50) | 0.07 |
| PetCO2/PaCO2 | 0.85 (0.75–0.91) | 0.81 (0.71–0.91) | 0.38 |
| RR, bpm | 17 (13–22) | 20 (15–25) | 0.003 |
| e, L/min | 7.3 (5.3–10.5) | 8.6 (6.5–11.7) | <0.001 |
| Vt, ml | 450 (369–541) | 490 (390–632) | 0.02 |
| Pressure support, cm H2O | 10 (9–12) | 10 (8–20) | 0.01 |
| PEEP, cm H2O | 13 (12–14) | 10.0 (5.0–10.0) | <0.001 |
| Extracorporeal blood flow, L/min | 2.57 (2–3) | 3.1 (3–3.2) | 0.02 |
| Sweep gas flow, L/min | 2 (0.73–3.35) | 1 (0.5–1.5) | <0.001 |
| ECMO LOS, d | 10 (6–18) | 10 (7–16) | 0.70 |
| Days to first weaning attempt | 8 (4–14) | 7 (5–12) | 0.47 |
| Weaning failure % | 42 | 37 | 0.582 |
| ICU LOS, d | 30 (16–38) | 20 (13–31) | <0.001 |
The complete list of the variables measured in the patients with success or failure are presented in Table E3. Among the variables that differed at baseline (before a weaning trial) between the success and failure groups in the physiological cohort, only the PetCO2/PaCO2 ratio (0.71 [0.63–0.83] vs. 0.91 [0.88–0.96]; P = 0.002) remained significantly associated with outcome when tested with univariate logistic regression (P = 0.021; Table E4). The ROC curve analysis of the PetCO2/PaCO2 ratio showed a significant area under the ROC curve (AUC): 0.87 (95% confidence interval [CI], 0.71–1), with the best cut-off point of the PetCO2/PaCO2 ratio for prediction of weaning success being ⩾0.84. This value correctly classified 86.4% of the patients and corresponded to a sensitivity of 91.7%, a specificity of 80%, and a positive likelihood ratio of 4.6.
When similar analyses were performed in the 638 patients of the clinical cohort, PetCO2/PaCO2 ratio remained the best predictor of success, with OR of 4.14 (95% CI, 1.32–12.2; P = 0.015) before initiation of weaning and OR of 13.1 (95% CI, 4–44.4; P < 0.001) for data collected at sweep gas flow of zero. The PaO2/FiO2 ratio at zero sweep flow was also associated with weaning outcome (OR, 1.06; 95% CI, 1.04–1.07; P < 0.001). ROC curve analysis of PetCO2/PaCO2 ratio in the clinical cohort showed an AUC of 0.58 (95% CI, 0.53–0.63), with the best cut-off point of the PetCO2/PaCO2 ratio for prediction of weaning success being ⩾0.83. This value correctly classified 58.2% of the patients and corresponded to a sensitivity of 54%, a specificity of 66%, and a positive likelihood ratio of 1.6.
A Cilley’s test (17) (the PaO2 measured at FiO2 of 1. 0 while on ECMO), total e, and rapid shallow breathing index before weaning and at zero sweep gas flow were not associated with weaning outcome. The PetCO2/PaCO2 ratio before the last weaning episode in the clinical cohort was 0.83 (0.72–0.91) for the 498 (96%) successfully weaned patients and 0.71 (0.62–0.82) for the 20 (4%) who were not successfully weaned (P < 0.01).
The main results of this study were: 1) The most frequent cause of weaning failure in the physiological cohort was the higher inspiratory effort required to clear the same amount of CO2; 2) PetCO2/PaCO2 ratio was significantly associated with weaning failure in the physiological cohort before initiation of weaning; and 3) in the clinical cohort of 638 patients, the two variables associated with outcome before initiation of weaning were PetCO2/PaCO2 ratio (OR, 4–13) and end PaO2/FIO2 ratio (OR, 1.06).
The rate of weaning failure from VV-ECMO in our physiological cohort (∼40%) was similar to the one we found in the clinical cohort and to those reported in several weaning studies (7, 8, 11). This finding suggests that variables other than the ones traditionally used are needed to improve the weaning outcome. A low PaO2/FiO2 ratio was the cause of interrupting the weaning attempt in 30% of our physiological cohort. Hypoxemia is related to significantly higher baseline venous admixture (P = 0.040), a variable usually not routinely measured in clinical practice. Excessive inspiratory effort was the predominant cause of failure in the physiological cohort (Figure E2). The mechanism appears quite clear: all patients increased e and inspiratory effort in response to a decrease in the volume of CO2 removed extracorporeally—an adaptation needed to maintain a PaCO2 nearly constant. This response is exactly the opposite of the one we observed experimentally when the CO2 is artificially removed: to maintain PaCO2 constant, the spontaneous e decreases, even to apnea (20). To eliminate a similar amount of CO2 per unit time, however, patients who eventually failed the weaning trial exerted an inspiratory effort that resulted in changes in pleural pressure exceeding 15 cm H2O (an arbitrary but already-acknowledged threshold for excessive effort) and double that of patients who passed the weaning test (21, 22). Notably the patients who failed weaning significantly increase the o2 to the natural lung throughout all the steps (Figure 2). The increase of effort was associated with significantly greater physiological Vd. Actually, the physiological Vd, to a greater extent than oxygenation, is the key variable associated with structural changes of lung parenchyma (23, 24).
The PetCO2/PaCO2 ratio was the unique (in the physiological cohort) and the best (in the clinical cohort) variable associated with weaning outcome. Indeed, the PetCO2/PaCO2 may be considered as a global bedside meter of the gas exchanger, theoretically equal to 1 under the ideal conditions met when alveolar Vd and venous admixture both equal 0. The lower the value, the worse the lung conditions (25, 26) (see online supplement). In a large database of patients with ARDS, the PetCO2/PaCO2 ratio was the variable best associated with impaired gas exchange, respiratory system mechanics, and computed tomography–assessed ARDS anatomy (26). Indeed, in our physiological cohort, it accurately reflected lung conditions leading to weaning failure, a status that was not detected by other variables normally used for weaning assessment. The value of the PetCO2/PaCO2 ratio as a predictive variable was confirmed in our retrospective analysis of the clinical cohort comprising 638 patients. It is important to comment on the fact that the AUC in the validation cohort was lower than that in the prospective cohort, despite similar threshold values of PetCO2/PaCO2 indicated by the ROC curve (0.84 vs. 0.83). The lower utility of using the exact point estimate to classify potential weaning success could be due to variability in the timing of calculation in the retrospective cohort, which may decrease the sensitivity and specificity. We would note that the OR of PetCO2/PaCO2 associated with weaning success was clinically significant in the validation cohort and an important consideration when considering probable success of weaning. The highest prediction was provided by the PetCO2/PaCO2 ratio at sweep gas flow of zero, with an OR of 13.
Taken together, the results suggest that even though the most common indication for commencing ECMO is oxygenation failure, the variables limiting weaning from ECMO relate to the lung’s inability to efficiently exchange carbon dioxide. In this context, a simple variable (PetCO2/PaCO2 ratio) could serve to assess weanability while the patient is still on ECMO or to stop the weaning trial early and return the patient to baseline ECMO to avert additional cardiovascular stress or potentially injurious inspiratory efforts. These adverse developments might otherwise remain unnoticed if success criteria were based solely on oxygen exchange efficiency or carbon dioxide tensions.
This study has several strengths. First, it prospectively studied a cohort of patients who underwent a standardized weaning protocol while monitoring gas exchange variables from both the natural and the extracorporeal lungs, as well as performed breath-by-breath measures of inspiratory effort. Second, the study included a large retrospective cohort of patients entered over a 10-year period who underwent comparable weaning trials. Third, the best predictor of weaning outcome that emerged from this analysis is a variable, the PetCO2/PaCO2 ratio, easily obtainable at the bedside in any critical care environment without the need of specialized or invasive equipment.
However, the study also presents several weaknesses. 1) In the physiological cohort we did not measure directly the o2 by the natural lung, which was only derived as the ratio between natural co2 and venous arterial CO2 content difference. 2) The computation of blood CO2 contents is problematic, as minimal error may produce large variations in the results (see online supplement for details). 3) The results of the regression model suffer from the limitations of a univariate analysis unadjusted for multiple potential confounders, as we aimed to demonstrate the utility of a single easily quantifiable bedside measurement rather than construct a more complex prognostic scoring system. 4) The clinical cohort was derived from a 10-year experience of a single center, and therefore the results may be affected by small variations in clinical practice over time. 5) Data that allow confident determination of the main cause for weaning failure, including assessment of right ventricular function, are lacking. Specifically, there were no physiological data recorded during the intermediate stages of the weaning trial itself, as the variables were collected before the weaning test (with sweep gas flow > 0) and after completion of weaning from the sweep gas flow, when patients were off ECMO. This lack of intervening data may have excluded patients who experienced early failures or had a trial of weaning initiated but did not achieve a sweep gas flow of zero. 6) The physiological variables that were collected for clinical reasons may refer to different time points after completion of the weaning trial. 7) Patients received assisted ventilation during weaning in both cohorts, which may not be the practice in other centers, and the amount of pressure support during weaning can affect the swings in esophageal pressure and inspiratory effort. These limitations should be considered when applying these findings. Despite these limitations, the study suggests that a variable not included in traditional failure criteria—but simply obtained at the bedside—can describe the lung’s efficiency in terms of gas exchange and simultaneously reflects inspiratory effort.
In conclusion, the primary reason for weaning failure from VV-ECMO is high effort to eliminate CO2. A higher PetCO2/PaCO2 ratio—a bedside measure of the overall gas exchanging performance of the natural lung—was associated with greater likelihood of weaning from VV-ECMO. Its calculation before weaning initiation may decrease the potentially detrimental effects of high inspiratory effort and increase the likelihood of successful weaning from VV-ECMO.
The authors thank all the ICU staff, nurses, and physicians for logistic support and contribution. They also thank Martin Sebastian Winkler, Francesca Rapetti, Giorgia Maiolo, Francesco Vasques, Eleonora Duscio, Jacopo Pasticci, Rosanna D’Albo, Martin Lier, and Francesco Cipulli for their valuable contribution in data collection.
| 1. | Zapol WM, Snider MT, Hill JD, Fallat RJ, Bartlett RH, Edmunds LH, et al. Extracorporeal membrane oxygenation in severe acute respiratory failure: a randomized prospective study. JAMA 1979;242: 2193–2196. |
| 2. | Combes A, Hajage D, Capellier G, Demoule A, Lavoué S, Guervilly C, et al.; EOLIA Trial Group, REVA, and ECMONet. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome. N Engl J Med 2018;378:1965–1975. |
| 3. | Tonna JE, Abrams D, Brodie D, Greenwood JC, Rubio Mateo-Sidron JA, Usman A, et al. Management of adult patients supported with venovenous extracorporeal membrane oxygenation (VV ECMO): guideline from the Extracorporeal Life Support Organization (ELSO). ASAIO J 2021;67:601–610. |
| 4. | Broman LM, Malfertheiner MV, Montisci A, Pappalardo F. Weaning from veno-venous extracorporeal membrane oxygenation: how I do it. J Thorac Dis 2018;10:S692–S697. |
| 5. | Grant AA, Hart VJ, Lineen EB, Badiye A, Byers PM, Patel A, et al. A weaning protocol for venovenous extracorporeal membrane oxygenation with a review of the literature. Artif Organs 2018;42:605–610. |
| 6. | Al-Fares AA, Ferguson ND, Ma J, Cypel M, Keshavjee S, Fan E, et al. Achieving safe liberation during weaning from VV-ECMO in patients with severe ARDS: the role of tidal volume and inspiratory effort. Chest 2021;160:1704–1713. |
| 7. | Gannon WD, Stokes JW, Bloom S, Sherrill W, Bacchetta M, Rice TW, et al. Safety and feasibility of a protocolized daily assessment of readiness for liberation from venovenous extracorporeal membrane oxygenation. Chest 2021;160:1693–1703. |
| 8. | Reeb J, Olland A, Pottecher J, Delabranche X, Schaeffer M, Renaud S, et al. Extracorporeal membrane oxygenation for acute respiratory distress syndrome after pneumonectomy. Ann Thorac Surg 2017;103:881–889. |
| 9. | Li X, Guo Z, Li B, Zhang X, Tian R, Wu W, et al. Extracorporeal membrane oxygenation for coronavirus disease 2019 in Shanghai, China. ASAIO J 2020;66:475–481. |
| 10. | Chaves RCF, Rabello Filho R, Timenetsky KT, Moreira FT, Vilanova LCDS, Bravim BA, et al. Extracorporeal membrane oxygenation: a literature review. Rev Bras Ter Intensiva 2019;31:410–424. |
| 11. | Seiler F, Trudzinski FC, Hörsch SI, Kamp A, Metz C, Flaig M, et al. Weaning from prolonged veno-venous extracorporeal membrane oxygenation (ECMO) after transfer to a specialized center: a retrospective study. J Artif Organs 2018;21:300–307. |
| 12. | Sen A, Callisen HE, Alwardt CM, Larson JS, Lowell AA, Libricz SL, et al. Adult venovenous extracorporeal membrane oxygenation for severe respiratory failure: current status and future perspectives. Ann Card Anaesth 2016;19:97–111. |
| 13. | Vasques F, Romitti F, Gattinoni L, Camporota L. How I wean patients from veno-venous extra-corporeal membrane oxygenation. Crit Care 2019;23:316. |
| 14. | Teijeiro-Paradis R, Tiwari P, Spriel A, Del Sorbo L, Fan E. Standardized liberation trials in patients with COVID-19 ARDS treated with venovenous extracorporeal membrane oxygenation: when ready, let them breathe! Intensive Care Med 2021;47:1494–1496. |
| 15. | Belliato M, Cremascoli L, Epis F, Ferrari F, Quattrone MG, Malfertheiner MV, et al. Carbon dioxide elimination during veno-venous extracorporeal membrane oxygenation weaning: a pilot study. ASAIO J 2021;67:700–708. |
| 16. | Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, et al.; ARDS Definition Task Force. Acute respiratory distress syndrome: the Berlin Definition. JAMA 2012;307:2526–2533. |
| 17. | Tonna JE, Abrams D, Brodie D, Greenwood JC, Rubio Mateo-Sidron JA, Usman A, Fan E. Management of adult patients supported with venovenous extracorporeal membrane oxygenation (VV ECMO): Guideline from the Extracorporeal Life Support Organization (ELSO). ASAIO J 2021;67:601-610. |
| 18. | Baydur A, Behrakis PK, Zin WA, Jaeger M, Milic-Emili J. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982;126:788–791. |
| 19. | Mauri T, Grasselli G, Suriano G, Eronia N, Spadaro S, Turrini C, et al. Control of respiratory drive and effort in extracorporeal membrane oxygenation patients recovering from severe acute respiratory distress syndrome. Anesthesiology 2016;125:159–167. |
| 20. | Kolobow T, Gattinoni L, Tomlinson TA, Pierce JE. Control of breathing using an extracorporeal membrane lung. Anesthesiology 1977;46: 138–141. |
| 21. | Mauri T, Yoshida T, Bellani G, Goligher EC, Carteaux G, Rittayamai N, et al.; PLeUral pressure working Group (PLUG—Acute Respiratory Failure section of the European Society of Intensive Care Medicine). Esophageal and transpulmonary pressure in the clinical setting: meaning, usefulness and perspectives. Intensive Care Med 2016;42:1360–1373. |
| 22. | Yoshida T, Brochard L. Esophageal pressure monitoring: why, when and how? Curr Opin Crit Care 2018;24:216–222. |
| 23. | Gattinoni L, Bombino M, Pelosi P, Lissoni A, Pesenti A, Fumagalli R, et al. Lung structure and function in different stages of severe adult respiratory distress syndrome. JAMA 1994;271:1772–1779. |
| 24. | Nuckton TJ, Alonso JA, Kallet RH, Daniel BM, Pittet JF, Eisner MD, et al. Pulmonary dead-space fraction as a risk factor for death in the acute respiratory distress syndrome. N Engl J Med 2002;346: 1281–1286. |
| 25. | Tusman G, Sipmann FS, Bohm SH. Rationale of dead space measurement by volumetric capnography. Anesth Analg 2012;114:866–874. |
| 26. | Bonifazi M, Romitti F, Busana M, Palumbo MM, Steinberg I, Gattarello S, et al. End-tidal to arterial PCO2 ratio: a bedside meter of the overall gas exchanger performance. Intensive Care Med Exp 2021; 9:21. |
* These authors contributed equally to this work and are co-first authors.
Supported by institutional funds and Sartorius (Otto-Brenner-Straße 20, 37079, Göttingen, Germany) for two unrestricted grants for research in respiratory medicine.
Author Contributions: Study concept and design: S.L., F.R., L. Gattinoni, L.C., J.J.M., and M.Q. Acquisition, analysis, or interpretation of data: S.L., B.S., L. Gattinoni, L.C., O.M., M. Busana, F.V., M. Bonifazi, M.M.M., L. Giosa, S.G., F.C., R.M., S.B., C.Z., D.H., L.-O.H., and M.G. First drafting of manuscript – writing committee: S.L., L. Gattinoni, F.R., L.C., B.S., J.J.M., and M.Q. Critical revision for important intellectual content and final approval of manuscript: F.R., L.C., J.J.M., M.Q., O.M., K.M., S.G., M. Busana, R.M., S.B., C.Z., F.V., M.M.M., M. Bonifazi, L. Giosa, F.C., D.H., L.-O.H., and M.G. Statistical analysis: S.L., L.C., and B.S. Administrative, technical, or material support: L.C., K.M., and O.M. Study supervision: L. Gattinoni.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Originally Published in Press as DOI: 10.1164/rccm.202201-0135OC on May 24, 2022
Author disclosures are available with the text of this article at www.atsjournals.org.
