ECMO Registry Data Show Progress in COVID-19

Around 60% of patients with COVID-19 who receive extracorporeal membrane oxygenation (ECMO) support survive to 90 days, according to registry data from 213 hospitals in 36 countries.

The analysis, published earlier this week in the Lancet, includes details on 1,035 COVID-19 patients tracked by the Extracorporeal Life Support Organization (ELSO), whose data set spans more than 100,000 people who’ve received ECMO for various indications.

The mortality rate—coming in at slightly below 40%—is good news, Ryan P. Barbaro, MD (University of Michigan, Ann Arbor), and colleagues say, since “early reports of patients with COVID-19 requiring ECMO suggested that mortality could be greater than 90%.” As reported this summer by TCTMD, results have indeed been mixed, though this hasn’t dashed hopes that ECMO could offer benefit to people infected with SARS-CoV-2.

ECMO’s more-general use in acute respiratory distress syndrome (ARDS) led to interest in its use for COVID-19, the researchers note. In May, the World Health Organization recommended that in settings with ECMO expertise, the therapy could be considered for COVID-19 patients with hypoxemia (PaO₂:FiO₂ < 50 mm Hg for 3 hours or < 80 mm Hg for more than 6 hours) despite lung protective ventilation.

Barbaro said the COVID-19 add-on to the ELSO registry, which he chairs, is meant to keep up with rapidly evolving information. “Because this [pandemic] was a place where we didn’t really know exactly what was going on, we had encouraged our centers to enter the data as it was happening. That’s a shift. And so even before cases were off of ECMO, people were entering them into the registry,” he said, adding that with their COVID-19 dashboard, “as soon it was entered, as soon as somebody struck the key, it came on the website.”

Being able to keep up to date was “important because at the beginning of this pandemic, [for] ECMO in particular, people were worried that outcomes were pretty terrible,” said Barbaro. “There was a legitimate question that people were saying: should we be using ECMO in this [setting]? Is COVID different enough that maybe we shouldn’t be using it just because of the disease?”

What this paper doesn’t address, he noted, is how strained healthcare systems should weigh the choice to expend the resources involved in providing ECMO. “What we were able to show with this and answer for people is that, in fact, the outcomes weren’t as concerning as the initial reports. And in fact they were a good bit better than what the initial reports suggested,” Barbaro pointed out.

ELSO Registry

In the ELSO registry, ECMO support—94% venovenous, 0.9% venoarterial, and 0.4% “other”—was initiated between January 16 and May 1, 2020, for the 1,035 COVID-19 patients (median age 49 years; 74% men). Their median body mass index (BMI) was 31 kg/m², and 70% had at least one comorbidity that preceded ECMO. Their illnesses included ARDS (79%), acute kidney injury (AKI; 29%), acute heart failure (5%), and myocarditis (2%). Around one-third (37%) had suspected or confirmed viral or bacterial co-infections, most commonly Staphylococcus aureus. Seventy percent had been transferred to an ELSO center from another hospital.

Prior to the start of ECMO, begun a median of 4 days after endotracheal intubation, the median PaO₂:FiO2 within the past 6 hours was 72 mm Hg. Common therapies were neuromuscular blockade (72%), prone positioning (60%), and noninvasive ventilatory support before intubation (59%), as well as vasoactive medications (60%) and COVID-19 therapies (76%). Around half were given hydroxychloroquine or chloroquine.

Median ECMO duration was 13.9 days, during which time 44% needed tracheostomy and 44% had renal replacement therapy. The most common complications were circuit change (15%), membrane lung failure (8%), central nervous syndrome hemorrhage (6%), and hemolysis (5%).

By the end of the study, 6% of the COVID-19 patients remained in the hospital, 37% died, and the rest were discharged to home or an acute rehabilitation center (30%), a long-term acute care center or unspecified location (10%), or another hospital (17%). Median length of hospital stay was 31.1 days for survivors and 16.0 days for the people who died.

Barbaro and colleagues assessed the primary outcome using time-to-event analysis. With this approach, the estimated cumulative incidence of in-hospital mortality within 90 days of ECMO initiation was 37.4% For patients with ARDS, 90-day mortality was 38.0%.

With a Cox proportional hazards model, venoarterial ECMO use was independently linked to in-hospital mortality (HR 1.89; 95% CI 1.20-2.97), as was age 70 years or older versus 16 to 39 years (HR 3.07; 95% CI 1.58-5.95). Other risk factors were AKI, chronic respiratory insufficiency, immunocompromised state, and pre-ECMO cardiac arrest. Higher PaO₂:FiO₂, on the other hand, was linked to lower mortality (HR 0.68 per doubling; 95% CI 0.57-0.81). Patient sex, BMI, and time between endotracheal intubation and ECMO had no bearing on survival.

“High mortality in the initial published experience led some clinicians and investigators to recommend withholding ECMO support in patients with COVID-19,” the investigators observe, adding, “Our findings provide provisional support for the use of ECMO in COVID-19-related acute hypoxemic respiratory failure.”

They caution, however, that these results can’t be extended to less-experienced centers, also calling for future research into what happens to these patients long term. Barbaro said this is challenging work to pursue, since many COVID-19 patients on ECMO were referred from other hospitals and after discharge will return to their usual source of care, making them hard to track.

‘Entering New Territory’

Asked what led to the dramatically better survival seen here versus in the pandemic’s early days, Barbaro offered two suggestions.

“When you’re entering new territory, sometimes you can be super conservative in your use of something, meaning you wait until it’s too late,” he said, noting that early ECMO trials outside of COVID-19 showed this phenomenon: by the time patients got ECMO, they were too sick for the therapy to help. An example is a 1979 JAMA paper focusing on patients with ARDS. Another factor may have been that the hospitals and healthcare systems were under more strain initially than in later phases of COVID-19.

For Barbaro, the experience of pulling together COVID-19 data under such difficult circumstances has been inspiring. “A community of people, while they were struggling with caring for patients in the heart of a pandemic and a lot of unknowns, were willing to rally together to change the way that they reported data, in order to help us all have a better understanding of how we can care for patients that are the sickest suffering from SARS-CoV-2,” he observed.

Going forward, the ELSO data might provide clarity on which patients are at risk of complications with ECMO and why, as well as on the role of COVID-19 therapies and immunomodulators, Barbaro said. There’s also room for research on what happens with pediatric patients, he added. “Fortunately there are not a lot, but there are children who have SARS-CoV-2 who ended up on ECMO support,” whether that’s more for respiratory symptoms or multisystem inflammatory syndrome.