• 
• 

Three Part Question

In [patients with an exacerbation of chronic obstructive pulmonary disease and respiratory acidosis], does [extracorporeal carbon dioxide removal] compared to [standard therapy] reduce the need for [invasive mechanical ventilation]?

Clinical Scenario

A patient arrives in your resuscitation bay in respiratory distress with a background of chronic obstructive pulmonary disease. After administering standard medical therapy, and controlled oxygen the patient remains in respiratory acidosis. Non-invasive ventilation is instituted, but the patient continues to be acidotic and deteriorates. The patient is prepared for rapid sequence induction and mechanical ventilation, but you wonder if there is an alternative?

Search Strategy

Search ((((((((Ecco2r[Title]) OR ((Carbon[Title]) AND Dioxide[Title])) OR Co2[Title]) OR Extracorporeal[Title]) OR Ecmo[Title])) AND (((Copd[Title]) OR ((((Acute[Title]) AND Respiratory[Title]) AND Distress[Title]) AND Syndrome[Title])) OR Novalung[Title])) AND (((((Novalung[Title]) OR Hemolung[Title]) OR iLa[Title]) OR (((Interventional[Title]) AND Lung[Title]) AND Assist[Title])) OR Decap[Title])) AND (((((((Ecco2r[Title]) OR ((Carbon[Title]) AND Dioxide[Title])) OR Co2[Title]) OR Extracorporeal[Title]) OR Ecmo[Title])) AND (((Copd[Title]) OR ((((Acute[Title]) AND Respiratory[Title]) AND Distress[Title]) AND Syndrome[Title])) OR Novalung[Title])). Human, English language, Adults 19 years+.
The databases Pubmed, Embase and Medline were searched from 1950 until July 2015.

Search Outcome

Pubmed=205, Embase=319, Medline=224. After removal of duplicates, three papers were identified from the search. A further nine papers were included which were not identified by the search criteria. The nomenclature regarding extracorporeal carbon dioxide removal makes identification of papers via search engines difficult.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Bartosik et al 2011 Ireland	• Use of the Novalung for self ventilating patients. • Two patients, one with connective tissue disease, and the other with COPD. • AV ECCO2R via Novalung was used in this case report.	Case report N=2.	The cases are presented as evidence of proof of principle in the use of ECCO2R in hypercapneic respiratory failure.	• Novolung was discontinued after four days and the patient later went on to receive lung transplantation. • No adverse events were reported for the COPD case. • The authors suggest that Novalung is simple and requires less skilled staff compared to ECMO. Some conclusions are based on the non-COPD case. • The authors claim that Novolung may be medium to lung term extracorporeal option for hypercapneic respiratory failure.	• Medical therapy received not stated. • pH and pCO2 are provided, but there is no data regarding respiratory rate, work of breathing or dyspnea scales. • The time line focuses on the initiation of ECCO2R and the effects on pH and pCO2, but does not provide context of standard medical therapy. • It is not stated whether NIV was used or not. However, NIV has been the standard of care for COPD respiratory acidosis for some years prior to this. • There is no discussion regarding anticoagulation, sweep gas flows.
Bonin et al 2013 Germany	• Single case report of a patient with an exacerbation of COPD and hypercapneic respiratory failure at risk of failure of NIV. • This study used VV ECCO2R (Hemolung).	Case study N=1.	The cases are presented as evidence of proof of principle in the use of ECCO2R in hypercapneic respiratory failure.	• The patient survived the episode of respiratory failure and went on to receive a lung transplant 31 days after cessation of ECCO2R. • This patient suffered an episode of bleeding as a result of over anticoagulation. • Provides a proof of principle to an emerging technology.	• There is no statement as why this particular case was selected from the Burki case series. Selection bias? • It is unclear from this report what the experience of the institution with ECMO. • This case represent the use of the most recent generation of ECCO2R technology using medium bore vena-venous a catheter. • There is minimal attempt to use the literature to support papers findings and conclusion. However, key points were base of use and assumption of lower risk of adverse event compared with IMV. • One author is an employee of the Hemolung manufacturer, the remaining authors declare no conflicts of interest. • The original Burki study had significant conflicts of interests.
Cardenas et al 2009 USA	• The case was a patient in her forties with COPD and asthma. The patient received IMV due to hypercapneic respiratory failure. • Presentation and symptoms are not mentioned. An ECMO machine was adapted to provide CO2 removal using a VV approach.	Case report N=1.	The cases are presented as evidence of proof of principle in the use of ECCO2R in hypercapneic respiratory failure.	• The patient was extubated two days after cessation of ECCO2R. The patient was discharged from hospital on day 13. • One episode of desaturation was reported which was attributed to large pH shift following ECCO2R initiation and omission of changes to IMV parameters. No bleeding, coagulopathy or other complications were noted.	• It is not clear in the abstract that this case involved the use of IMV as part of the management of a COPD exacerbation or why NIV was not used. It was only evident halfway through this paper that the patient received intubation and IMV and not NIV. • At the time of publication, this was the earliest care report referring to COPD that the search criteria identified. • The paper on closer analysis does not fit the defined PICO question, but due to the title and abstract construction, this was not evident in the inclusion criteria. • Little supporting evidence is considered external to this case report regarding the conclusions reached. • First use of ECCO2R for the institution involved in this case report. Given the type of ECMO machine used, this was likely an institution with full ECMO capability and a high level of experience, despite this novel application.
Cole et al 2014 UK	• The patient had an exacerbation of COPD and was failing to respond to NIV. • This study used VV ECCO2R (Hemolung).	Case report N=1.	Proof of principle benefit based on risks of IMV for NIV failure.	• The patient survived to discharge from hospital and had not been re-admitted within a thirty day period. • The intervention was well tolerated, and allowed the patient to discontinue NIV without any deleterious effect. • There is no discussion regarding any adverse events during the therapy.	• The institution that provided the care was a commissioned ECMO centre. • This represent use of the most recent generation of ECCO2R technology using medium bore vena-venous catheters. Expert centre in the use of ECMO. • The author presents a strong theoretical case for the potential benefit for ECCO2R, but does not attempt to assert this case as proof of efficacy. The need for a randomised controlled trial is asserted. • Conclusions are very circumspect and emphasise the theoretical benefits and do not attempt to assert a proof of efficacy. • One author has received travel expenses from Alung (manufacturer of Hemolung) but had received no direct payments. One author is also on the medical advisory board of Hemolung.
Mani et al 2011 India and Germany.	• Two patients with an exacerbation of COPD and hypercapneic respiratory failure at risk of failure of NIV. • This study used VV ECCO2R (Hemolung).	Case report N=2.	The cases are presented as evidence of proof of principle in the use of ECCO2R in hypercapneic respiratory failure.	• Both patients survived to hospital discharge, one was re-admitted at thirty days. In both cases, intubation was avoided. • Both patients developed thrombocytopenia, treatment was not required and reversed following cessation of ECCO2R. • An argument is put forward that due to the mortality associated with NIV failure in this context, avoiding intubation with ECCO2R offers theoretical advantages.	• It is not clear why these particular patients were chosen from the case series by Burki. This may represent positive reporting bias as four other patients in the NIV plus ECCO2R group in the original study died within the follow up period. • It is unclear from these two reports what the experience of these two institutions was. Were these centres with previous experience of ECCO2R or ECMO? • There is minimal attempt to use the literature to support papers findings and conclusion. However, key points were base of use and assumption of lower risk of adverse event compared with IMV. • One author is an employee of the Hemolung manufacturer. • The original Burki study held significant conflicts of interests.
Burke et al 2013 Germany and India.	• Group 1 N=7 COPD patients at risk of NIV failure. • Group 2 N=2 failure to wean from NIV. • Group 3 N=11 patients that had failed to wean from IMV. • The whole patient group had COPD exacerbation as the primary diagnosis. • This study used VV ECCO2R (Hemolung).	Case series N=20.	• The paper describes a pilot study in the use of ECCO2R in patients in various states of hypercapneic respiratory failure. • The emphasis of the study was of feasibility and safety. • No clear primary or secondary outcomes or objectives are stated.	• Group 1. All patients avoided IMV. 3 died within 30 days of ECCO2R cessation. One patient spent less than 30min of ECCO2R as the catheter clotted. 6 patients experienced reduction in dyspnea, increased pH and reduced pCO2. 1 patient was able to stop NIV while on ECCO2R. • Group 2. One patient was able to stop NIV while on ECCO2R the other was able to have reduced support. Both experienced reduced dyspnea scores and improved pCO2, and pH. Both patients were alive at 30 days post ECCO2R on intermittent NIV. • Group 1 and 2, decrease in pCO2 was statistically significant (p value .003). The increase in pH was also statistically significant (p value .002) • ECCO2R was able improve pCO2, and pH in patients receiving NIV and suffering a COPD exacerbation who appeared to be failing NIV therapy. • Reduced dyspnea scores following ECCO2R. • One patient died from retroperitoneal haemorrhage as a result of the venous catheterisation procedure. One pneumothorax also occurred. • Apart from the one death through haemorrhage, bleeding episodes were related to anticoagulation.	• Inherent biases of a case series study design. • Exclusion criteria not stated. • No information regarding COPD medical managment. • No pre-defined outcome measures. • First author is a consultant and medical director for ALung, who make the Hemolung machine. • A number of ALung employees provided editorial assistance and data inspection. • This study was part of data collection for European Union regulatory application process. • The combination of patients in receipt of NIV and IMV makes interpretation difficult. • This study at best presents proof of concept.
Del Sorbo et al 2015 Italy.	• Case patients were had an exacerbation of COPD in receipt of NIV and were deemed at risk of failure to respond to NIV. Matched controls. • Control patients were taken from data of previous studies at the same institution and were included if they fitted the same inclusion criteria for ECCO2R. • This study used VV ECCO2R.	Case series and historical control. Cases N=25. Historical matched controls N=21.	Population focus specifically on patients with COPD and hypercapneic respiratory failure deemed at risk of failing NIV. Very clearly stated primary endpoint of 28 day prevalence of intubation in the case group. Study also aimed to look at safety of ECCO2R.	• Significant difference in pH and pCO2 between groups one hour post ECCO2R, 21% difference in pCO2 (p value .0003 for pH and .012 for pCO2), • The effect of the above changes lead to a difference in respiratory rate of 22% between groups after one hour of ECCO2R. • 12% intubation in the ECCO2R + NIV group vs 33% in the NIV. ARR of 21%, RR 64%. NNT 4.6. • Hospital mortality 8% rate ECCO2R + NIV group vs 33% in the NIV group. ARR education of 30% and RRR of 79%. OR 0.14. NNT 3.3. • No difference in length of hospital stay. • 52% complications from ECCO2R, the most serious being a retroperitoneal haemotoma. All patients with ECCO2R that were intubated were related to the failure of the ECCO2R system. These were largely due to clots in the ECCO2R system and the development of respiratory acidosis. • Reduction in the risk of need for IMV was statistically significant with a p value of .47, and a narrow confidence interval of 0.07-0.98. • Differences in hospital morality was statistically significant p value .034, wide CI.	• Excluded patients that failed to consent could have been used as controls. • Temporally, the control group may not be representative of cases. • Over time, NIV failure rate may have reduced. • The confidence interval of the primary outcome is narrow despite the overall sample size. • The first author is a consultant and medical director for ALung, who make the Hemolung machine. • ALung provided machines and circuits for the study centers. • A number of ALung employees provided editorial assistance and data inspection.
Kluge et al 2012 Germany and India.	• Population focus on patients with hypercapneic respiratory failure who were considered to be failing to respond or deteriorating following application of NIV. No specification of the disease aetiology. 66% of case group had COPD. • This study used AV pumless ECCO2R.	Case control. Cases N= 21. Controls N=21 (all controls had failed NIV and received IMV.	rms feasibility, effectiveness and safety used as opposed to a specific pre-defined outcome	• Need for intubation 10% in ECCO2R group. • No significant difference in ITU or hospital length of stay (p value .53 and .32). • No statistically significant difference in 28 day or six month mortality (24% versus 19% and 33% versus 33%). • 33% of patients experienced complications related to ECCO2R, these centered on bleeding, and possibly related more to arterial canulation rather than venous canula.	• Both groups had at least one patient with a long length of stay that will distort results with such small numbers of patients. • Using a control group who all received IMV is a significant confounder. • A high number of cases in this study were on a lung transplant waiting list. This may have determined more aggressive treatment. • This population may not represent the majority of COPD patients who develop hypercapneic acidosis. • Multiple authors have ties to Novalung, the manufacturer of iLA.
Braune and Kluge 2013 Germany	A review of ECCO2R and its application to patients with COPD.	Narrative review.	A question as such is not posed, but a focused statement regarding the potential indications, safety and effectiveness in regards to COPD is made.	• A reduction in risk of complications from ECCO2R compared to ECMO are asserted. However, the authors also argue that risks or ECCO2R must be weighed against the benefits. • Further studies are needed to identify subsets of patients where patients may benefit from ECCO2R. Proof of concept has been established, but randomised controlled trials are needed. • Use of ECCO2R in avoidance of IMV in COPD patients is the most radical innovation, this ideal scenario being the combination of NIV and ECCO2R where NIV alone is failing. • ECCO2R technology has seen marked improvements and innovations in the last few year. • The aim of ECCO2R is to reduce or avoid the complications associated with IMV.	• The majority of evidence considered is low level, a combination of case series, controls, case reports and expert opinion. • As the COPD literature is limited in regard to ECCO2R, there is a degree of extrapolation from ARDS literature. • A systematic data base search was not carried out, nor was there in any use of inclusion/exclusion criteria applied. • There is little attempt to critique any of the referenced literature, no critical appraisal tools were applied. • There is little use of statistics to support assertions, as there are only two studies identified in the reference list and these are case control or case series. • The findings are borne out through expert opinion and proof of principle rather than statistical inference. • Many of the quoted studies and case reports involve centers with experience of ECMO and although the COPD patients themselves can probably be generalised, the experience and expertise of the centers can not. • A reduction in risk of complications from ECCO2R compared to ECMO are asserted. However, the authors also argue that risks or ECCO2R must be weighed against the benefits. • Kluge has ties to Novolung based on conflicts of interest.
Lund and Frederspiel 2013 Germany	A review of ECCO2R and its application to patients with COPD.	Narrative review.	A question as such is not posed, but a focused statement regarding the potential indications, safety and effectiveness in regards to COPD is made.	• There is substantial increase in mortality and morbidity associated with NIV failure resulting in the use of IMV. • Up to 50% of CO2 can be removed with ECCO2R and sweep gas flow can be titrated to respiratory rate. • There is insufficient evidence to initiate ECCO2R where NIV is not failing. • Risks or ECCO2R over ECMO are reduced, but must be acknowledged. • Current evidence prevents conclusions regarding mortality or length of stay. • Reduced need for IMV is argued. • Randomised controlled trials are also needed on the use of ECCO2R in the weaning of COPD patients from IMV and NIV failure.	• The ideal studies for a review following a hierarchy of evidence such as randomised controlled trials, systematic reviews do not exist. • There is no attempt to systematically search the evidence base or provide inclusion and exclusion criteria. • This paper looks at a number of case reports and a case series and also makes some extrapolations from ECCO2R applied to ARDS patients. Some effort is also made that addresses the issue of NIV failure and the associated mortality in this group. • A large proportion of the review literature focused on NIV and the associated risks of failure, and on historical aspects of ECMO and ECCO2R. All the relevant papers regarding ECCO2R and NIV to that date were considered. • No critical appraisal tools were applied to the referenced literature. • There is little use of data to support assertions. • The findings are borne out through expert opinion and proof of principle rather than statistical analysis. • Both authors are employees of Alung, a VV ECCO2R system manufacturer. Lund appears in three other papers.
Terragni et al 2012 Italy.	Technical aspects of ECCO2R and application in ARDS, COPD and as a bridge to transplantation.	Narrative review.	A review of the technical aspects of ECCO2R and its clinical application.	• Combining ECCO2R with ultra protective ventilation may be a promising therapeutic option in the future. • ECCO2R is new potential strategy in the management of COPD exacerbation. • Randomised controlled trials are needed.	• The majority of evidence considered is low level, a combination of case series, controls, case reports and expert opinion. The paper considers ECCO2R literature in general, although does have a focused section on COPD. • The key papers that have been published to date were considered. The COPD/ECCO2R references did not include papers identified in this papers systematic search. • A systematic data base search was not carried out, nor was there in any use of inclusion/exclusion criteria applied. • Many of the quoted studies and case reports involve centers with experience of ECMO and although the COPD patients themselves can probably be generalised, the experience and expertise of the centers can not.

Comment(s)

The evidence for the use of extracorporeal carbon dioxide removal (ECCO2R) in the context of chronic obstructive pulmonary disease (COPD) exacerbation and a failed response to non-invasive ventilation (NIV) is small and recent. This is emphasised in the identified narrative reviews which cite a handful of the same papers. They extrapolate extensively from the literature surrounding ECCO2R in the context of acute respiratory distress syndrome (ARDS). In the case of ARDS, Fitzgerald et al (2014) a systematic review only elicited two randomised controlled trials (RCT). The bulk of the literature lies within the lower hierarchy of evidence in the form of case series and case-controls. This low level evidence may be an appropriate beginning in the developing literature, given the very recent nature of the application of ECCO2R to patients with COPD and acute exacerbations. In view of the cost of new technology and its implementation, proof of principle is a minimum requirement before expensive, large scale and multi centre RCT can be contemplated. Conflicts of interests are evident within much of this literature, including; funding, employment, editorial assistance and provision of equipment from industry. With regard to ECCO2R and its effectiveness in the removal of CO2, there is little doubt. The NIV and ARDS literature show consistently that the technique is effective. Using this removal of CO2 as the surrogate endpoint does not necessarily translate into a meaningful patient centred outcome and its use must be treated with caution. There is a strong theoretical basis that reducing or avoiding the need for invasive mechanical ventilation (IMV) in patients who have failed NIV may translate to an improvement in morbidity and mortality. This argument combined with the effectiveness of CO2 removal with ECCO2R, is an important bench to bed side translation of research and a pertinent research question. At present, the theoretical advantage of ECCO2R seems to be the primary argument for its use, implying cause and effect. Within the limits of the available level of evidence, there is support for the notion that ECCO2R can reduce the need for IMV in this population. Taken at face value, Del Sorbo et al (2015) found a significant treatment effect. However, it is important not to automatically assume cause and effect. Chandra et al (2012) noted increased mortality in patients who required IMV when NIV failed to stabilise COPD patients. Conversely, patients who received IMV from the outset had an improved outcome compared to patients who failed NIV later and then received IMV. There may be other factors at play with regard to morbidity and mortality in this group of patients which are not immediately apparent. It is possible that reducing IMV in this group of patients may not provide the hoped-for mortality benefit, and other confounders may be present. VV-ECCO2R (veno-venous) is likely to supersede the initial pump-less arterio-venous (AV) system which was designed to be a less invasive alternative to ECMO. The equivalent CO2 removal capacity of VV-ECCO2R, without the risks of large bore arterial cannulation, make AV-ECCO2R of little additional value. This is particularly pertinent when balancing the risks of ECCO2R with IMV, and could potentially switch the benefits to harm. The additional benefit of VV-ECCO2R its increased utility outside of specialist ECMO centres. If future research is to establish a firm mortality benefit, it will not be possible to transport unstable patients who are not responding to NIV around the country to regional ECMO centres. It seems that if there is to be a future for ECCO2R, it will be in its integration into the typical therapies currently available in the average intensive care unit. There is insufficient evidence to make assumptions about the impact of ECCO2R upon mortality or morbidity. Whilst proof of principle has been established, this does not mean to say that the reduction of NIV failure should continue to be accepted as a surrogate for reduced mortality. The risks of major bleeding and even death from haemorrhage should not be downplayed. Exchanging the risk of mortality from IMV with that of major bleeding and infection should not be considered advantageous. Using a comparable therapeutic method that is expensive, invasive and not widely available is not in the best interests of the patient. There is insufficient data to draw conclusions as to whether ECCO2R is a safer alternative to IMV. There are signals in the current body of evidence that ECCO2R may reduce the need for IMV in patients with COPD with an exacerbation which does not respond to NIV. It is unclear whether there is a mortality benefit or if there is a reduction in morbidity.

Clinical Bottom Line

An RCT is required which is adequately powered to detect statistically significant differences in the need for invasive ventilation, mortality and morbidity between groups. It is difficult to justify the risks of AV-ECCO2R versus a VV approach. VV-ECCO2R should be the mode of choice, particularly in the context of use outside of the domain of an ECMO centre. There is insufficient evidence to recommend the routine use of ECCO2R in any population. In centres where a system of training and clinical governance is in place, ECCO2R is a potential option in individual cases.

References

1. Bartosik, W. Egan, J. Wood, A. The Novalung interventional lung assist as a bridge to lung transplantation for self-ventilating patients- Initial expereince. Interactive cardiovascular and thoracic surgery. 2011; 198-200.
2. Bonin, F. Sommerwerck, U. Lund, L. Teschler, H. Avoidance of intubation during acute exacerbation of chronic obstructive pulmonary disease for a lung transplant candidate using extracorporeal carbon dioxide removal with the Hemolung. The journal of thoracic and cardiovascular surgery. 2013; 43-44.
3. Cardenas, V. Cardenas, V. Lynch, J. Ates, R. Miller, L. Zwishenberger, J. Venovenous carbon dioxide removal in chronic obstructive pulmonary disease: experience in one patient. American society for artificial internal organs. 2009; 420-422.
4. Cole, S. Barrett, N. Glover, G. Langrish, C. Extracorporeal carbon dioxide removal as an alternative to endotracheal intubation for non-invasive ventilation failure in acute exacerbation of COPD. Journal of the intensive care society. 2014; 344-346.
5. Mani, R.K., Schmidt, W., Lund, L.W., and Herth, F.J.F. Respiratory dialysis for avoidance of intubation in acute exacerbation of COPD. ASAIO 2013; 675-678.
6. Burki, N.K., Mani, R.K., Herth, F.J.F., Schmidt, W., Teschler, H., Bonin, F., Becker, H., Randerath, W.J., Stieglitz, S., Hagmeyer, L., Priegnitz, C., Pfeifer, M., Blaas, S.H., Putensen, C., Theuerkau A novel extracorporeal CO(2) removal system: results of a pilot study of hypercapnic respiratory failure in patients with COPD. Chest 2013; 678-686.
7. Del Sorbo, L., Pisani, L., Filippini, C., Fanelli, V., Fasano, L., Terragni, P., Dell’Amore, A., Urbino, R., Mascia, L., Evangelista, A., Antro, C., D’Amato, R., Sucre, M.J., Simonetti, U., Persico, P Extracorporeal Co2 removal in hypercapnic patients at risk of noninvasive ventilation failure: a matched cohort study with historical control. Critical care medicine. 2015; 120-127.
8. Kluge, S., Braune, S.A., Engel, M., Nierhaus, A., Frings, D., Ebelt, H., Uhrig, A., Metschke, M., Wegscheider, K., Suttorp, N., and Rousseau, S. Avoiding invasive mechanical ventilation by extracorporeal carbon dioxide removal in patients failing noninvasive ventilation. Intensive care medicine. 2012; 1632-1639.
9. Braune, S.A. and Kluge, S. Extracorporeal Lung Support in Patients with Chronic Obstructive Pulmonary Disease. Minerva Anestesiolgica 2013; 934-943.
10. Lund, L.W. and Federspiel, W.J. Removing extra CO2 in COPD patients. Current Respiratory Care Reports. 2013; 131-138.
11. Terragni, P., Maiolo, G., and Ranieri, V.M. Role and potentials of low-flow CO(2) removal system in mechanical ventilation. Current Opinion in Critical Care. 2012; 93-98.