It is estimated that over 350 adult patients suffer from severe, but potentially reversible, respiratory failure in the UK each year. The mortality rate for such patients is very high and has only improved marginally in the majority of centres over the last 20 years[
1,
2] Current management uses intermittent positive pressure ventilation (IPPV). The airway pressures and oxygen concentrations required to maintain adequate blood gases are often very high in patients with severe respiratory failure, and this combination of barotrauma, volutrauma and oxygen toxicity can prevent lung recovery. An alternative treatment, extracorporeal membrane oxygenation (ECMO), uses cardio-pulmonary bypass technology to temporarily provide gas exchange to patients with severe, but potentially reversible, respiratory failure. During ECMO, ventilator settings can be reduced, and such 'lung-rest' allows the lungs to recover. There is currently no good evidence from randomised controlled trials (RCTs) to compare ECMO against conventional management for important clinical outcomes.
Patients are usually considered for ECMO when they have such severe disease that they continue to deteriorate despite maximal optimum 'conventional' treatment. For the purposes of this protocol, conventional will be defined as any treatment which relies on the patient's lungs to provide gas exchange. Conventional treatment may therefore include inhaled nitric oxide and prone ventilation[
3‐
5], as well as the more usual types of positive pressure ventilation. The use of ECMO to support
neonatal patients with severe respiratory failure has been rigorously evaluated in an RCT[
6,
7]. The neonatal ECMO RCT convincingly demonstrated the effectiveness of ECMO in improving patient survival without severe disability. Neonatal ECMO in the UK is now a supra-regional service receiving central funding. The use of ECMO as it is currently practised in older children[
8], and adults[
9] is more controversial, and has yet to be evaluated in an RCT in the UK.
Previous studies
A review of the literature was carried out to identify all studies relevant to adult ECMO. Only two RCTs have been reported[
1,
10], both in the United States but they used such different approaches that they have not been combined as a formal meta-analysis. Each is detailed below, followed by the recent non-experimental evidence.
An RCT of adult ECMO was conducted by the US National Institutes of Health (NIH)[
1], in the early days of extra-corporeal support in the 1970s. Survival in both groups was very poor (around 10%), and no difference was shown in survival between the conventional and ECMO treated groups. There were a number of important differences in the perfusion and ventilation techniques used during this trial compared to those used today. Firstly, veno-arterial (VA) rather than veno-venous (VV) perfusion was used, and this was thought to be responsible for the high incidence of pulmonary micro-thrombosis and fibrosis seen in the lungs of the ECMO patients (due to reduced pulmonary blood flow). Secondly, patients were anti-coagulated to such a degree that severe bleeding occurred. Thirdly, high pressure ventilation was continued during ECMO resulting in continued barotrauma and volutrauma[
11,
12]. Finally, the mean duration of ventilation prior to ECMO in the NIH ECMO trial was over 9 days, whereas it is now well-recognised that after 7 days of high pressure ventilation with high fraction of inspired oxygen (FIO
2) the lungs only have limited powers of recovery[
13].
More recently there has been an RCT of the related technique of extra-corporeal carbon dioxide removal (ECCO
2R)[
10]. This showed no difference between ECCO
2R and conventional treatment. Again there were numerous differences in the clinical and perfusion protocols between this trial and those in widespread use in the majority of centres currently[
14]. Firstly, the experimental arm of the trial used low flow ECCO
2R in a group of patients who had severe lung disease, which warranted higher flow ECMO with full support of oxygenation and carbon-dioxide removal. This was demonstrated by the need to increase the airway pressure in the ECCO
2R group half-way through the study. The reliance on the patient's lungs to provide oxygenation, especially at such high airway pressures, also eliminated any possibility for lung rest. Also, despite the involvement of one of the team in the 1970s NIH ECMO trial, in which VA ECMO was used with very small numbers in each centre (<5), the ECCO
2R programme in this trial was not well developed prior to the study (as the team had only provided ECCO
2R to sheep and one patient prior to starting the trial). The high incidence of bleeding and thrombotic complications reported in this study may attest to this inexperience. In addition, the conventional treatment used in the trial was Pressure Controlled Inverse Ratio Ventilation (PCIRV) using a computer controlled algorithm. The results of this treatment showed 44% survival compared to expected survivals of < 20% in other similar series of patients[
2]. Despite this, survival in the ECCO
2R group was the same as the 'conventional' group. The success of the PCIRV protocol in this study has led to the wide adoption of the technique within 'conventional' ventilatory management with survival of 66% for patients with
moderate to severe respiratory failure (mean Murray score 2.8, mean ratio between the oxygen tension in the arterial blood and the fraction of inspired oxygen (PaO
2/FIO
2) 88 mmHg)[
15]. Unfortunately no other authors have been able to duplicate the PCIRV results of Morris et al. for patients with
severe progressive respiratory failure.
Because the two trials described above have little relevance to the ECMO regimens used in the majority of centres worldwide, the only relevant evidence consists of observational studies. By the nature of their design, the information they provide is potentially biased, and must therefore be viewed with caution.
Recent case series of patients with similar degrees of respiratory failure to the eligibility criteria for the second trial suggest survival rates without ECMO of 18% to 44%[
1,
10]. compared to rates of up to 66% with high flow ECMO (including full support of oxygenation and lung rest), provided by experienced teams principally in the USA, UK and Germany[
9,
13,
14].
In a cohort study of the first 50 adult patients to receive ECMO for respiratory support at Glenfield Hospital, Leicester, UK, patients had severe respiratory failure as shown by the mean pre-ECMO Murray Lung Injury Score of 3.4 (SD 0.5) and PaO
2/FIO
2 ratio of 65 mmHg (SD 36.9). They were referred for ECMO with severe respiratory failure caused by either the Acute Respiratory Distress Syndrome (ARDS) or with pneumonia. The overall survival rate was 66%[
9].
For the reasons outlined above, it is impossible to reach firm conclusions from the above experimental and observational data regarding the clinical effectiveness or cost-effectiveness of VV high flow ECMO for respiratory failure in adults. The recent evidence from observational studies does, however, suggest that ECMO could potentially be a highly useful treatment in these patients. The case selection and treatment protocols used during ECMO are now well defined by the international Extracorporeal Life Support Organization (ELSO), and the only team using ECMO in adults consistently in the UK has built up clinical expertise[
9].
It is not possible to further define the safety and efficacy of ECMO as a treatment without a rigorous trial. The procedure has received a Cii categorisation (safety and/or efficacy not yet fully established; procedure requires a fully controlled evaluation) from the UK Safety and Efficacy Register of the New Interventional Procedures of the Medical Royal Colleges (SERNIP). Additionally a situation of equipoise currently exists, whereby clinicians can see the potential benefits of ECMO, but do not have enough evidence to make an informed choice as to the best treatment for their patient.
The aim of the present trial is therefore to assess whether for patients with severe, but potentially reversible, respiratory failure, ECMO will increase the rate of survival without severe disability by six months post randomisation and will be cost effective from the viewpoints of the NHS and society, compared to conventional ventilatory support.