In adults with acute hypoxemic respiratory failure receiving mechanical ventilation, does further reduction in tidal volumes, facilitated by extracorporeal carbon dioxide removal, improve 90-day mortality compared with conventional low tidal volume ventilation?
In this randomized clinical trial that included 412 adults, 90-day mortality was 41.5% in the extracorporeal carbon dioxide removal group and 39.5% in the standard care group, a difference that was not statistically significant.
Among patients with acute hypoxemic respiratory failure, the use of extracorporeal carbon dioxide removal to facilitate lower tidal volume ventilation, compared with conventional low tidal volume ventilation, did not significantly reduce 90-day mortality.
In patients who require mechanical ventilation for acute hypoxemic respiratory failure, further reduction in tidal volumes, compared with conventional low tidal volume ventilation, may improve outcomes.
To determine whether lower tidal volume mechanical ventilation using extracorporeal carbon dioxide removal improves outcomes in patients with acute hypoxemic respiratory failure.
Design, Setting, and Participants
This multicenter, randomized, allocation-concealed, open-label, pragmatic clinical trial enrolled 412 adult patients receiving mechanical ventilation for acute hypoxemic respiratory failure, of a planned sample size of 1120, between May 2016 and December 2019 from 51 intensive care units in the UK. Follow-up ended on March 11, 2020.
Participants were randomized to receive lower tidal volume ventilation facilitated by extracorporeal carbon dioxide removal for at least 48 hours (n = 202) or standard care with conventional low tidal volume ventilation (n = 210).
Main Outcomes and Measures
The primary outcome was all-cause mortality 90 days after randomization. Prespecified secondary outcomes included ventilator-free days at day 28 and adverse event rates.
Among 412 patients who were randomized (mean age, 59 years; 143 [35%] women), 405 (98%) completed the trial. The trial was stopped early because of futility and feasibility following recommendations from the data monitoring and ethics committee. The 90-day mortality rate was 41.5% in the lower tidal volume ventilation with extracorporeal carbon dioxide removal group vs 39.5% in the standard care group (risk ratio, 1.05 [95% CI, 0.83-1.33]; difference, 2.0% [95% CI, −7.6% to 11.5%]; P = .68). There were significantly fewer mean ventilator-free days in the extracorporeal carbon dioxide removal group compared with the standard care group (7.1 [95% CI, 5.9-8.3] vs 9.2 [95% CI, 7.9-10.4] days; mean difference, −2.1 [95% CI, −3.8 to −0.3]; P = .02). Serious adverse events were reported for 62 patients (31%) in the extracorporeal carbon dioxide removal group and 18 (9%) in the standard care group, including intracranial hemorrhage in 9 patients (4.5%) vs 0 (0%) and bleeding at other sites in 6 (3.0%) vs 1 (0.5%) in the extracorporeal carbon dioxide removal group vs the control group. Overall, 21 patients experienced 22 serious adverse events related to the study device.
Conclusions and Relevance
Among patients with acute hypoxemic respiratory failure, the use of extracorporeal carbon dioxide removal to facilitate lower tidal volume mechanical ventilation, compared with conventional low tidal volume mechanical ventilation, did not significantly reduce 90-day mortality. However, due to early termination, the study may have been underpowered to detect a clinically important difference.
ClinicalTrials.gov Identifier: NCT02654327
Acute hypoxemic respiratory failure is a leading cause of admission to intensive care units (ICUs) and is associated with significant mortality and long-term morbidity for survivors, as well as considerable resource implications for health care systems.1 A significant proportion of patients affected by acute hypoxemic respiratory failure will meet the diagnostic criteria for acute respiratory distress syndrome (ARDS).2 Invasive mechanical ventilation after tracheal intubation is often used as a life-saving intervention to maintain adequate gas exchange, but is known to contribute to the overall morbidity and mortality of this condition.3
One of the few interventions shown to reduce mortality in patients with acute hypoxemic respiratory failure and ARDS is ventilation with a lung-protective strategy aiming for a tidal volume of 6 mL/kg predicted body weight and a plateau pressure less than or equal to 30 cm H2O in patients.4 However, even when using lung-protective invasive mechanical ventilation, lung hyperinflation and injury can still occur.5 Reducing tidal volumes further may result in respiratory acidosis, which can cause further adverse effects, such as pulmonary hypertension and altered cardiac function. Extracorporeal gas exchange, including extracorporeal carbon dioxide removal (ECCO2R), can facilitate mechanical ventilation with even lower tidal volumes because it supports the removal of carbon dioxide that accumulates in this setting.6,7 The feasibility of ECCO2R in patients with acute hypoxemic respiratory failure due to ARDS has recently been demonstrated.8
The primary objective of the REST trial was to determine whether lower tidal volume ventilation facilitated by ECCO2R compared with standard care in patients with acute hypoxemic respiratory failure decreased mortality 90 days after randomization.9
Trial Design and Oversight
This was a multicenter, randomized, allocation-concealed, open-label, pragmatic clinical trial. After randomization, patients, clinical care clinicians, and researchers were unblinded due to the complex nature of the intervention. The trial was coordinated by the Northern Ireland Clinical Trials Unit and was sponsored by Belfast Health and Social Care Trust. The study design has been published9 and the trial protocol and statistical analysis plan are provided in Supplement 1 and Supplement 2. The protocol was approved by research ethics committees in England, Wales, Northern Ireland (16/SC/089), and Scotland (16/SS/048). The National Institute for Health Research in the UK convened an independently chaired (and majority independent) trial steering committee and an independent data monitoring and ethics committee. The study was conducted in accordance with Good Clinical Practice guidelines, local regulations, and the ethical principles described in the Declaration of Helsinki. Written informed consent from patients or agreement from their surrogates was obtained, keeping with regional regulations.
The trial was conducted in 51 adult, general ICUs within the National Health Service across the UK. Patients 16 years or older who were admitted to a participating ICU were eligible for inclusion if they had an acute and potentially reversible cause of moderate to severe hypoxemic respiratory; were receiving invasive mechanical ventilation using at least 5 cm H2O of positive end-expiratory pressure (PEEP); and were within 48 hours of onset of hypoxemia, defined as a ratio of the partial pressure of oxygen in arterial blood to the fractional inspired concentration of oxygen (Pao2/Fio2) of less than 150 mm Hg. Exclusion criteria included receiving invasive mechanical ventilation for more than 7 days, contraindication to limited systemic anticoagulation with heparin, untreated pulmonary embolism, pleural effusion or pneumothorax, or acute respiratory failure fully explained by left ventricular failure or fluid overload. Other reasons for exclusion are detailed in the trial protocol in Supplement 1.
After consent was obtained, eligible patients were randomized. Randomization concealment was achieved by use of an automated online or telephone-centralized 24-hour randomization facility. Patients were randomized to receive lower tidal volume ventilation with ECCO2R or lung protective ventilation alone in a 1:1 ratio using a computer-generated schedule with variable block sizes of 4, 6, and 8, stratified by recruitment center. If randomized to the ECCO2R group, it was recommended to commence within 8 hours of randomization.
In patients assigned to receive ECCO2R, a dual-lumen catheter was inserted percutaneously into a central vein using ultrasonography guidance. Venovenous ECCO2R was then commenced using intravenous heparin as systemic anticoagulation to prevent circuit thrombosis. The pump speed was increased to achieve the maximum possible blood flow (typically 350-450 mL/min), sweep gas flow was increased to 10 L/min to maximize carbon dioxide removal, and concomitantly tidal volumes were reduced incrementally, aiming for a tidal volume less than or equal to 3 mL/kg predicted body weight. The intervention was continued for at least 48 hours, after which patients were weaned from ECCO2R, as per the trial manual provided in Supplement 3, when patients demonstrated signs of clinical improvement and improvement in the degree of hypoxemia. ECCO2R was to be used for a maximum of 7 days as part of the study protocol. An online educational package for catheter insertion and device management was provided to all sites.
For patients randomized to receive standard care, it was recommended that patients received mechanical ventilation using a tidal volume of 6 mL/kg predicted body weight with PEEP set based on the ARDSNetwork trial.4 In addition, in keeping with UK guidelines,10 patients in both the intervention and control groups could receive neuromuscular-blocking drugs,11 prone positioning,12 or referral for consideration of extracorporeal membrane oxygenation (ECMO).13
The primary outcome was all-cause mortality 90 days after randomization. Secondary clinical outcome measures were tidal volume at day 2 and day 3, ventilator-free days at day 28, duration of invasive mechanical ventilation in survivors, need for ECMO up to day 7, mortality at day 28, and adverse event rate. All outcomes were reported from the time of randomization. Prespecified clinical outcome measures are listed in eTable 1 in Supplement 4. The outcomes not reported in this article will be reported separately. A cost-effectiveness analysis is also planned, as described in the protocol in Supplement 1. Duration of critical care and hospital length of stay were defined as outcomes for the cost-effectiveness analysis. Data on physiological parameters by treatment group were also collected up to day 7.
A sample size of 1120 patients was determined to provide 90% power to show an absolute difference of 9% in 90-day mortality, assuming a control group mortality rate of 41%.14 This postulated effect size was estimated from a previous trial on the use of lung protective ventilation,4 which demonstrated a 9% reduction in mortality in patients with hypoxaemic respiratory failure secondary to ARDS with a 50% reduction in tidal volume (from 12 to 6 mL/kg predicted body weight).15 Therefore, we hypothesized that a similar relative reduction in tidal volume would result in a 9% difference in mortality. The sample size calculation did not take a group sequential trial design into account.
Patients were analyzed according to their randomization group. For the primary outcome and other dichotomous outcomes, risk ratios and percent point differences with 95% CIs were calculated. The primary outcome of 90-day mortality was analyzed using a χ2 test and a secondary analysis using a log-binomial regression adjusted for age, sequential organ failure assessment (SOFA)16 score, and baseline Pao2/Fio2 ratio was also carried out. Plateau pressure was planned to be included as a variable in the adjusted analysis, but because it was missing in a substantial number of patients this was not possible. A post hoc sensitivity analysis using generalized estimating equations was used to account for possible clustering of observations within participating centers. There was no imputation for missing data. Continuous outcomes were compared between the 2 groups using analysis of variance/analysis of covariance, adjusting for other covariates where appropriate. Time-to-event outcomes were analyzed by survival methods and reported as hazard ratios with 95% CIs. The proportionality assumption was tested using the Schoenfeld test. Length of stay outcomes were compared using the Wilcoxon rank sum test. A prespecified sensitivity analysis was also performed for the primary outcome excluding the first 2 intervention group patients at each site to address potential learning effects. A per-protocol analysis was carried out for the secondary outcome of tidal volume at days 2 and 3 (ie, including those who were receiving ECCO2R on day 2 and 3 in the intervention group). We performed prespecified subgroup analyses using 99% CIs. Log-binomial regression was used with interaction terms (treatment group × subgroup).
One interim analysis for the primary outcome was planned before the recruitment of 560 patients. A post hoc conditional power analysis was carried out estimating the power given the observed data up to termination and then assuming varying differences between 2% and 10% for the remainder of the data that were to be observed. Because of the potential for type I error due to multiple comparisons, findings for analyses of secondary end points should be interpreted as exploratory. Analysis was conducted using Stata/SE, version 15.1 (StataCorp). Statistical significance was defined using a 2-sided test with α = .05.
During a recruitment pause for investigation of a serious adverse event (SAE) (fatal intracranial hemorrhage), the planned interim analysis was undertaken and the independent data monitoring and ethics committee recommended that the trial be stopped due to futility (given that even under optimistic assumptions the trial was unlikely to demonstrate a significant benefit for the intervention) and subsequent feasibility to continue the trial. There were no formal stopping rules for futility and the decision to stop the study was not based on a formal calculation of futility. The decision to stop was based on the opinion of the data monitoring and ethics committee based on all available information, including data from the interim analysis, feasibility of future recruitment, and a conditional power analysis. The conditional power analysis to detect a difference between the groups, assuming the patients remaining to be recruited to achieve the planned sample size met the assumptions of the original sample size, was 44%. Safety was not cited by the data monitoring and ethics committee as a reason for stopping the trial. This decision was accepted by the trial steering committee and agreed by the study sponsor and the trial was stopped on February 11, 2020.
From May 2016 to December 2019, a total of 7071 patients from 51 centers were screened for eligibility and, after applying the exclusion criteria, 412 (6%) participants were recruited (eTable 2 in Supplement 4). The patients were followed up until March 11, 2020. One patient was randomized twice in error, 2 patients were lost to follow-up, and 4 patients withdrew consent for confirmation of vital status. As a result, 405 participants (200 in the intervention group and 205 in the standard care group) were included in the final analysis of the primary outcome (Figure 1). The most common reasons for exclusion were contraindication to systemic anticoagulation, a do-not-attempt-resuscitate order in place, imminent treatment withdrawal, and invasive mechanical ventilation for more than 7 days. A total of 1866 patients (28%) who were screened were excluded for other reasons, for which the most common reason was either the patient’s clinical condition rapidly improved or deteriorated. The baseline characteristics of the 2 groups were well balanced prior to randomization and typical of patients with moderate to severe acute hypoxemic respiratory failure requiring ICU care (Table 1).
There was no significant difference in mortality between the groups. The 90-day mortality rate was 41.5% (83 of 200) in the intervention group and 39.5% (81 of 205) in the standard care group (risk ratio [RR], 1.05 [95% CI, 0.83-1.33]; difference, 2.0% [95% CI, −7.6% to 11.5%]) (Table 2 and Figure 2). The RR was similar after adjustment for age, SOFA score, and Pao2/Fio2 ratio (RR, 1.12 [95% CI, 0.90-1.40]) and in a per-protocol analysis for the primary outcome of the group who initiated ECCO2R (Table 2). To address a potential learning effect with the intervention, a sensitivity analysis was performed excluding the first 2 patients randomized to receive the intervention at each site (Table 2). These findings were consistent with those of the primary analysis. Treatment × subgroup interactions were not significant with respect to the presence of ARDS, requirement for vasopressors, severity of hypoxemia or hypercapnia, plateau and driving pressures, Acute Physiology and Chronic Health Evaluation II score, and volume of ECCO2R by site (eFigure 1 in Supplement 4). The percentage of missing data for the primary analysis of the primary outcome was 1.7%.
The secondary outcomes are presented in Table 2. There were significantly fewer ventilator-free days at day 28 in the intervention group (7.1 [95% CI, 5.9-8.3] vs 9.2 [95% CI, 7.9-10.4] days; mean difference, −2.1 days [95% CI, −3.8 to −0.3]; P = .02). There was no significant between-group difference in duration of ventilation, need for ECMO at day 7, mortality at 28 days, or duration of ICU or hospital stay.
Additional Secondary Outcomes and Intervention Fidelity
Of the 202 patients allocated to the intervention group, 186 patients (92%) received ECCO2R after randomization, with a mean (SD) duration of ECCO2R of 4 (2) days. One patient in the standard care group received nonprotocol ECCO2R for 2 days. A total of 50 patients (28%) were successfully weaned from ECCO2R and it was stopped due to receiving 7 days of treatment in 33 patients (18%). ECCO2R was discontinued for safety reasons in 14 patients (8%), the need for ECMO in 12 patients (7%), and withdrawal of active medical treatment or death in 28 patients (16%) (eTable 3 in Supplement 4).
Patients randomized to receive ECCO2R had a lower tidal volume than those randomized to receive standard care at day 2 (4.5 [95% CI, 4.3-4.8] vs 6.5 [95% CI, 6.3-6.7] mL/kg; mean difference, 2.0 mL/kg [95% CI, 1.7-2.3]) and day 3 (4.4 [95% CI, 4.1-4.6] vs 6.7 [95% CI, 6.4-7.0] mL/kg; mean difference, 2.3 mL/kg [95% CI, 2.0-2.7]). In patients receiving ECCO2R on day 2 and 3, tidal volume was lower than in the standard care group at day 2 (4.2 [95% CI, 4.0-4.4] vs 6.5 [95% CI, 6.3-6.7] mL/kg; mean difference, 2.4 mL/kg [95% CI, 2.0-2.7]) and day 3 (3.8 [95% CI, 3.6-4.0] vs 6.7 [95% CI, 6.4-7.0] mL/kg; mean difference, 2.9 mL/kg [95% CI, 2.5-3.3]) (Figure 3A; eTable 4 in Supplement 4).
Patients in the intervention group, compared with the standard care group, had a lower Pao2/Fio2 ratio on day 2 (147.8 [95% CI, 140.4-155.1] vs 161.1 [95% CI, 153.3-169.0]; mean difference, 13.3 mm Hg [95% CI, 2.6-24.1]) and on day 3 (147.9 [95% CI, 140.9-154.9] vs 167.0 [95% CI, 158.6-175.4]; mean difference, 19.1 mm Hg [95% CI, 8.2-30.1]) after randomization (eFigure 2A and eTable 4 in Supplement 4). Patients in the intervention group had higher PEEP than patients in the control group (Figure 3B; eTable 4 in Supplement 4). Plateau pressure was lower in the intervention group on day 2 (23.5 [95% CI, 22.6-24.3] vs 25.7 [95% CI, 24.9-26.6]; mean difference, 2.3 cm H2O [95% CI, 1.1-3.4]) and on day 4 (22.2 [95% CI, 21.2-23.1] vs 23.7 [95% CI, 22.6-24.8]; mean difference, 1.6 cm H2O [95% CI, 0.1-3.0]) after randomization (Figure 3C; eTable 4 in Supplement 4). Driving pressure was lower in the intervention group than in the control group from day 2 to 5 following randomization (Figure 3D; eTable 4 in Supplement 4). Total respiratory rate was higher in the intervention group than the control group from day 2 to 4 following randomization (day 2: 26.6 [95% CI, 25.8-27.3] vs 24.6 [95% CI, 23.9-25.3]; mean difference, 2.0 breaths per minute [95% CI, 0.9-3.0]; day 3: 27.8 [95% CI, 26.9-28.7] vs 24.4 [95% CI, 23.6-25.2]; mean difference, 3.4 breaths per minute [95% CI, 2.2-4.6]; day 4: 27.0 [95% CI, 26.0-28.1] vs 24.4 [95% CI, 23.5-25.4]; mean difference, 2.6 breaths per minute [95% CI, 1.2-4.0]; eFigure 2B and eTable 4 in Supplement 4). Minute ventilation was lower in the intervention group than in the control group from day 1 following randomization (eFigure 2C and eTable 4 in Supplement 4). Paco2 was higher from day 2 following randomization (eFigure 2D and eTable 4 in Supplement 4) and pH was lower in the intervention group than the standard care group following randomization (eFigure 2E and eTable 4 in Supplement 4). The rate of carbon dioxide removal is shown in eTable 4 in Supplement 4.
Patients in the intervention group were more likely to receive ventilation with a mandatory mode of ventilation to day 7 (59 [39.1%] vs 32 [18.9%]; difference, 20.1% [95% CI, 10.4-29.9]) on day 7), received more neuromuscular blockade from day 2 following randomization (110 [55.6%] vs 92 [44.2%]; difference, 11.3% [95% CI, 1.7%-21.0%] on day 2), and received ventilation less frequently in the prone position on days 1 (11 [5.5%] vs 29 [13.9%]; difference, −8.4% [95% CI, −14.0% to −2.8%]) and 2 (18 [9.1%] vs 27 [13.0%]; difference, −3.9% [95% CI, −10.0% to 2.2%]) following randomization (eTable 5 in Supplement 4).
Adverse event rates are presented in Table 3. Adverse events were more common in the intervention group. Eighty patients experienced SAEs (62 [31%] in the intervention group and 18 [9%] in the standard care group). Twenty-one patients experienced 22 SAEs related to the study device (eTable 6 in Supplement 4). There were 12 events defined as intracranial hemorrhage (9 of which were defined as SAEs), all of which occurred in the intervention group. Of these SAEs, 5 were considered to be at least possibly related to the intervention by the site investigator (3 patients had an intracerebral hemorrhage and 2 patients had a subarachnoid hemorrhage). An additional 4 SAEs were considered unlikely to be related to the intervention by the site investigator (1 patient had an intracerebral hemorrhage and 3 had hemorrhagic changes on brain imaging). There were 21 events defined as bleeding at other sites, 7 of which were defined as SAEs, with 6 occurring in the intervention group and 1 in the control group. Of those SAEs in the intervention group, 4 were considered to be at least possibly related to the intervention by the site investigator (airway bleeding, hemothorax in a patient with chest trauma, bleeding from a venous hemodialysis catheter, and a hematoma at an attempted vascular access site). The additional 2 SAEs were considered unlikely to be related to the intervention by the site investigator (upper gastrointestinal bleeding and pharyngeal bleeding following reintubation). The event in the control group was an episode of rectal bleeding.
Conditional Power Analysis
Post hoc conditional power analysis for mortality showed a conditional power of 4% for a 2% effect size, 8% for a 4% effect size, 17% for a 6% effect size, 31% for an 8% effect size, and 48% for a 10% effect size.
In this UK multicenter, randomized clinical trial that was stopped early due to futility, tidal volume reduction during invasive mechanical ventilation facilitated by ECCO2R, compared with standard care, in patients with acute hypoxemic respiratory failure did not reduce mortality at 90 days.
The aim of supportive care with invasive mechanical ventilation in patients with acute hypoxemic respiratory failure during the past 20 years has moved away from targeting normal gas exchange to limiting ventilator-induced lung injury.3,4 A secondary analysis of the ARMA trial suggested there may be no safe threshold for tidal volumes with in vivo data providing biological plausibility for the benefit of further reduction in tidal volumes.15,17 In the study, a reduction in mean tidal volumes of 2.0 mL/kg at day 2 and 2.3 mL/kg at day 3 were achieved, from a prerandomization tidal volume of 6.6 mL/kg, with significant reduction in tidal volumes to day 7, which were associated with significant reductions in plateau and driving pressures. It was mandated that the intervention was applied for at least 48 hours to ensure an effective “dose” of lower tidal volumes, although it is possible that a longer duration of ECCO2R, with greater tidal volume reduction, may have been required to demonstrate an effect because higher intensities of invasive mechanical ventilation have been shown to be associated with increased risk of death in a time-dependent fashion.18 Duration of ECCO2R in the study was limited to less than 7 days due to regulations associated with use of the device, and the intervention was discontinued in 33 patients for this reason. It is unknown whether the results would have changed had these 18% of patients in the intervention group received longer ECCO2R treatment. The primary aim of the trial was to lower tidal volumes facilitated by extracorporeal carbon dioxide removal. Permissive hypercapnia was tolerated to enable tidal volume reduction.19 There was a lower Pao2/Fio2 ratio, higher respiratory rate, and greater hypercapnia and respiratory acidosis in the intervention group, although these effects were modest. As a result, harmful effects associated with these physiological consequences could have contributed to the lack of clinical benefit. Furthermore, that the minute ventilation was reduced indicates that the increase in respiratory rate is unlikely to have offset the reduction in ventilator-induced lung injury achieved with tidal volume reduction. The effect of a larger reduction in ventilator-induced lung injury on outcome remains unknown. Lung-protective ventilation has also been demonstrated to improve outcomes in patients with acute hypoxemic respiratory failure without ARDS,20 so the aim was to include a broad cohort that would reflect the general population of critically ill patients who may benefit. A systematic review concluded that although evidence was limited, ECCO2R was feasible and had been shown to facilitate further reduction in tidal volumes with the potential to mitigate ventilator-induced lung injury and improve outcomes in patients with more severe hypoxemia.21 This work informed the use of a Pao2/Fio2 ratio of less than 150 mm Hg as the qualifying level of hypoxemia for this study population.22
After adjustment for age, degree of hypoxia, and organ dysfunction, the primary outcome was unchanged. Furthermore, results of subgroup analyses did not suggest that the effects of the intervention were modified by any of the variables investigated. Although results of subgroup analyses showed that other baseline characteristics associated with ventilator-induced lung injury did not have an effect on outcome, it remains unknown whether a different population might benefit from ECCO2R. Enrichment strategies to identify a population that may be more likely to benefit are needed for future trials of ECCO2R.23,24
Five patients in the current study were reported to have intracranial hemorrhage related to the intervention. This incidence is comparable to data from previous trials of ECMO in severe acute respiratory failure.13,25 A review of changes in Paco2, presence of thrombocytopenia or coagulopathy, and the degree of therapeutic anticoagulation and blood pressure was performed in these patients, but it was not possible to identify a clear mechanism for these events. Patients with severe hypoxemic respiratory failure have an increased risk of intracranial hemorrhage, with recent data reporting a background rate of intracranial hemorrhage in patients with severe hypoxemic respiratory failure to be approximately 8% to 10%, although this was substantially increased in patients receiving ECMO.26,27
The study has several limitations. First, only 6% of screened patients were included in the study, which may limit the generalizability of the results. Second, 17 patients (8.4%) did not receive the intervention as randomized, which could have diluted the effect in the intervention group, although a per-protocol analysis of patients who received the intervention did not change the outcome. Third, most of the sites were naive to the intervention before the study commenced. Although an extensive educational package and training program addressing catheter insertion and maintenance of the device was put in place at all sites, it is possible that practical inexperience with the intervention may have negatively affected the outcomes in the intervention group. Volume-outcome relationships have been previously reported with ECMO.28 In an attempt to address a potential learning effect, a sensitivity analysis excluding the first 2 patients randomized to the intervention group at each site was undertaken that showed no notable change to the primary outcome, and there was no significant difference in a subgroup analysis between sites that recruited more or fewer than 10 patients to the trial. Fourth, other aspects of care were not standardized in each group because this was a pragmatic trial, and clinicians were free to treat patients as they would normally. The use of the intervention was associated with longer use of neuromuscular-blocking drugs and less prone positioning. Although the difference in the use of neuromuscular-blocking drugs is unlikely to have modified outcome, the less frequent use of prone positioning could have affected the outcome in the intervention group, albeit the absolute difference in the use of prone positioning between the groups was relatively small.11,12 Fifth, it is possible that the trial was underpowered to detect a clinically important difference, particularly because the trial was stopped before recruitment of the planned sample size was achieved. Sixth, due to the complexity of the intervention, blinding to the clinicians or patients was not possible, which could have resulted in performance bias.
In patients requiring mechanical ventilation for acute hypoxemic respiratory failure, lower tidal volume ventilation facilitated by extracorporeal carbon dioxide removal, compared with standard care, did not result in a reduction in mortality at 90 days. However, due to early termination, the study may have been underpowered to detect a clinically important difference.
Corresponding Author: Daniel F. McAuley, MD, Wellcome-Wolfson Institute for Experimental Medicine, School of Medicine, Dentistry and Biomedical Science, Queen’s University of Belfast, 97 Lisburn Rd, Belfast BT9 7AE, United Kingdom (firstname.lastname@example.org).
Accepted for Publication: July 25, 2021.
Published Online: August 31, 2021. doi:10.1001/jama.2021.13374
Correction: This article was corrected on December 15, 2021, to correct an error in the excluded individuals in Figure 1 and in the labels in Figure 2 in Supplement 4. The information in these figures is now presented correctly.
Author Contributions: Drs McNamee and McAuley and Ms McDowell had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: McNamee, Gillies, Barrett, Perkins, Young, Bentley, Harrison, Brodie, McDowell, McAuley.
Acquisition, analysis, or interpretation of data: McNamee, Barrett, Perkins, Tunnicliffe, Bentley, Harrison, Brodie, Boyle, Millar, Szakmany, Bannard-Smith, Tully, Agus, McDowell, Jackson, McAuley.
Drafting of the manuscript: McNamee, Gillies, Barrett, Millar, McDowell, McAuley.
Critical revision of the manuscript for important intellectual content: McNamee, Barrett, Perkins, Tunnicliffe, Young, Bentley, Harrison, Brodie, Boyle, Millar, Szakmany, Bannard-Smith, Tully, Agus, Jackson, McAuley.
Statistical analysis: McNamee, Barrett, Boyle, McDowell.
Obtained funding: McNamee, Gillies, Barrett, Perkins, Young, Agus, McAuley.
Administrative, technical, or material support: McNamee, Gillies, Barrett, Perkins, Tunnicliffe, Bentley, Brodie, Boyle, Jackson, McAuley.
Supervision: McNamee, Barrett, Young, Bannard-Smith, McAuley.
Conflict of Interest Disclosures: Dr McNamee reported receiving grants from the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) Programme during the conduct of the study and speaking fees from Baxter outside the submitted work. Dr Barrett reported receiving grants from ALung Technologies during the conduct of the study and grants from ALung for a separate study using the investigational device outside the submitted work. Dr Perkins reported receiving grants from NIHR HTA Programme and the NIHR Applied Research Collaboration West Midlands during the conduct of the study. Dr Young reported receiving grants from the NIHR during the conduct of the study. Dr Harrison reported receiving grants from the NIHR during the conduct of the study. Dr Brodie reported receiving grants from ALung Technologies; personal fees from Baxter, Xenios, and Abiomed; and nonfinancial support from Hemovent outside the submitted work and being the chair of the executive committee of the International ECMO Network. Dr Boyle reported receiving grants from the Northern Ireland Health and Social Care Research and Development Agency to undertake a clinical trial of extracorporeal carbon dioxide removal, grants from the NIHR HTA Programme for a clinical trial of extracorporeal carbon dioxide removal (13/143/02), and nonfinancial support from ALung Technologies in provision of equipment and consumables to undertake a clinical trial of ECCO2R. Dr Agus reported the Northern Ireland Clinical Trials Unit receiving funds from the NIHR HTA Programme for its involvement during the conduct of the study. Dr McDowell reported receiving grants from the NIHR HTA Programme during the conduct of the study. Dr Jackson reported receiving grants from the NIHR HTA Programme during the conduct of the study. Dr McAuley reported receiving grants from the NIHR HTA Programme during the conduct of the study and personal fees from Bayer for consultancy for treatment of acute respiratory distress syndrome (ARDS), GlaxoSmithKline for consultancy for treatments of ARDS and being an educational seminar speaker, Boehringer Ingelheim for consultancy for treatment of ARDS, Novartis for consultancy for treatment of COVID-19, Eli Lilly for consultancy for treatment of COVID-19, and Vir Biotechnology as a member data monitoring and ethics committee and grants from NIHR as an investigator in ARDS and COVID-19 studies, Wellcome Trust as an investigator in ARDS and COVID-19 studies, Innovate UK as an investigator in ARDS and COVID-19 studies, the Medical Reserve Corpse as an investigator in ARDS studies, and Northern Ireland Health and Social Care Research and Development Division as an investigator in ARDS and COVID-19 studies outside the submitted work; having a patent issued by Queen's University Belfast for a novel treatment for inflammatory disease (US8962032); and being co-director of research for the intensive care society and NIHR/Medical Reserve Corps Efficacy and Mechanism Evaluation Programme. No other disclosures were reported.
Funding/Support: This project was funded by the National Institute for Health Research Health Technology Assessment Programme (13/143/02) with additional funding from the Health Research Board. Dr Perkins is supported by the National Institute for Health Research Applied Research Collaboration (ARC) West Midlands. The trial was supported by the UK Intensive Care Society, the United Kingdom Critical Care Research Group, and the International ECMO Network. This study was also supported by the Irish Critical Care Clinical Trials Network at the University College Dublin, which is funded by the Health Research Board of Ireland (CTN 2014-012).
Role of the Funder/Sponsor: The ECCO2R devices, catheters, and consumables were provided free of charge by the manufacturer, ALung Technologies. ALung Technologies had no role in the study design or in the study conduct, data analysis, or data interpretation. This trial was commissioned and funded by the National Institute for Health Research. The sponsor and funder approved the design of the study and monitored the conduct of the study. They played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The views expressed are those of the authors and not necessarily those of the National Health Service, the National Institute for Health Research, or the Department of Health.
Group Information: Members of the REST trial are listed in Supplement 5.
Data Monitoring and Ethics Committee: Marion Campbell, PhD, University of Aberdeen, Aberdeen, United Kingdom; Tom Clutton-Brock, MD, University of Birmingham, Birmingham, United Kingdom; Taylor Thompson, MD, Harvard Medical School, Boston, Massachusetts; Niall Ferguson, MD, University of Toronto, Toronto, Ontario, Canada.
Trial Steering Committee: William McGuire, MD (University of York, York, United Kingdom); Barry Williams, LLB (patient and lay representative; Dorset, United Kingdom); John Laffey, MD (National University of Ireland, Galway, Ireland); and Alistair Nichol, PhD (University College Dublin, Dublin, Ireland).
Clinical Trials Unit Team: Christine McNally; Glenn Phair, MSc; Mark Wilson, MSc; Sorcha Toase, MSc; Danielle Logan, PhD; Kelly Green, BA(hons); Judith McCrory, BSc; Natasha Loughran, BSc; Barbara Thompson, PhD; Lauren Holmes, PhD; Aine Carson; Jeanette Mills, BSc; Loren McGinley-Keag, BSc (Northern Ireland Clinical Trials Unit, Royal Hospitals, Belfast, United Kingdom).
Meeting Presentation: The study was presented at the Critical Care Reviews Conference; January 21, 2021.
Data Sharing Statement: See Supplement 6.
Additional Contributions: We thank all the patients who participated in the trial as well as the medical and nursing staff in participating centers who cared for patients and collected data. We are grateful for the support of the Intensive Care National Audit and Research Centre (ICNARC) for providing APACHE II data for sites that participate in ICNARC’s Case Mix Programme. The authors acknowledge the support of the Northern Ireland Clinical Research Network and the National Institute for Health Research Clinical Research Network.
et al; LUNG SAFE Investigators; ESICM Trials Group. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries. JAMA
. 2016;315(8):788-800. doi:10.1001/jama.2016.0291PubMedGoogle ScholarCrossref
et al; ARDS Definition Task Force. Acute respiratory distress syndrome: the Berlin Definition. JAMA
. 2012;307(23):2526-2533.PubMedGoogle Scholar
A; Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med
. 2000;342(18):1301-1308. doi:10.1056/NEJM200005043421801PubMedGoogle Scholar
et al; Pressure- and Volume-Limited Ventilation Strategy Group. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. N Engl J Med
. 1998;338(6):355-361. doi:10.1056/NEJM199802053380603PubMedGoogle ScholarCrossref
et al; International ECMO Network (ECMONet). Extracorporeal carbon dioxide removal for lowering the risk of mechanical ventilation: research questions and clinical potential for the future. Lancet Respir Med
. 2018;6(11):874-884. doi:10.1016/S2213-2600(18)30326-6PubMedGoogle ScholarCrossref
VM; European Society of Intensive Care Medicine Trials Group and the “Strategy of Ultra-Protective lung ventilation with Extracorporeal CO2 Removal for New-Onset moderate to severe ARDS” (SUPERNOVA) investigators. Feasibility and safety of extracorporeal CO2
removal to enhance protective ventilation in acute respiratory distress syndrome: the SUPERNOVA study. Intensive Care Med
. 2019;45(5):592-600. doi:10.1007/s00134-019-05567-4PubMedGoogle ScholarCrossref
et al. Protective ventilation with veno-venous lung assist in respiratory failure: a protocol for a multicentre randomised controlled trial of extracorporeal carbon dioxide removal in patients with acute hypoxaemic respiratory failure. J Intensive Care Soc
. 2017;18(2):159-169. doi:10.1177/1751143716681035PubMedGoogle ScholarCrossref
et al; CESAR trial collaboration. Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial. Lancet
. 2009;374(9698):1351-1363. doi:10.1016/S0140-6736(09)61069-2PubMedGoogle ScholarCrossref
RG; ARDS Clinical Trials Network. Tidal volume reduction in patients with acute lung injury when plateau pressures are not high. Am J Respir Crit Care Med
. 2005;172(10):1241-1245. doi:10.1164/rccm.200501-048CPPubMedGoogle ScholarCrossref
et al. Lung protective mechanical ventilation and two year survival in patients with acute lung injury: prospective cohort study. BMJ
. 2012;344:e2124. doi:10.1136/bmj.e2124PubMedGoogle Scholar
et al; LUNG SAFE Investigators and the ESICM Trials Group. Patterns and impact of arterial CO2 management in patients with acute respiratory distress syndrome: insights from the LUNG SAFE study. Chest
. 2020;158(5):1967-1982. doi:10.1016/j.chest.2020.05.605PubMedGoogle ScholarCrossref
et al. Association between use of lung-protective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress syndrome: a meta-analysis. JAMA
. 2012;308(16):1651-1659. doi:10.1001/jama.2012.13730PubMedGoogle ScholarCrossref
et al. Extracorporeal carbon dioxide removal for patients with acute respiratory failure secondary to the acute respiratory distress syndrome: a systematic review. Crit Care
. 2014;18(3):222. doi:10.1186/cc13875PubMedGoogle ScholarCrossref
et al. Lower tidal volume strategy (≈3 ml/kg) combined with extracorporeal CO2 removal versus ‘conventional’ protective ventilation (6 ml/kg) in severe ARDS: the prospective randomized Xtravent-study. Intensive Care Med
. 2013;39(5):847-856. doi:10.1007/s00134-012-2787-6PubMedGoogle ScholarCrossref
et al; SUPERNOVA investigators (European Society of Intensive Care Medicine trials group) and for the International ECMO Network (ECMONet). Determinants of the effect of extracorporeal carbon dioxide removal in the SUPERNOVA trial: implications for trial design. Intensive Care Med
. 2019;45(9):1219-1230. doi:10.1007/s00134-019-05708-9PubMedGoogle ScholarCrossref
et al; EOLIA Trial Group, REVA, and ECMONet. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome. N Engl J Med
. 2018;378(21):1965-1975. doi:10.1056/NEJMoa1800385PubMedGoogle ScholarCrossref
et al. Intracranial hemorrhage and early mortality in patients receiving extracorporeal membrane oxygenation for severe respiratory failure. Semin Thromb Hemost
. 2018;44(3):276-286. doi:10.1055/s-0038-1636840PubMedGoogle Scholar