'Exceptional' ECMO Outcomes in COVID-19 Reflect Early Choices

Respiratory and survival outcomes in the year after receiving extracorporeal membrane oxygenation (ECMO) for severe COVID-19 were excellent in a highly selected group treated at one New York City hospital early in the pandemic.

Of 30 such patients who underwent venovenous (VV) ECMO from March 10 to May 1, 2020 after conventional mechanical ventilation failed, 27 (90%) survived to discharge and 26 (86.7%) survived to 1 year post-cannulation.

All survivors were discharged home or to short-term rehabilitation without ventilator dependence, including three who had been maintained on VV-ECMO for more than 100 days.

Only one of the 27 required supplemental oxygen on discharge, reported Deane E. Smith, MD, of NYU Langone Health in New York City, and colleagues in the Annals of Thoracic Surgery.

Survivors who had testing 1 year after leaving the hospital had recovered lung function, with a median percent predicted forced expiratory volume in 1 second (FEV₁) of 100%. One patient had progressive post-COVID-19 pulmonary fibrosis and required lung transplantation but did well afterward and was discharged home on room air on postoperative day 7.

The researchers highlighted the "exceptional" early survival results that were sustained at 1 year. "We learned that lung recovery was actually possible. When the pandemic started, that was not clear," said Smith in a statement.

Brian Mitzman, MD, of the University of Utah in Salt Lake City, who helped manage ECMO patients at NYU Langone during the height of the pandemic, who was not part of the study, agreed that the results were "great," but encouraged taking them with a few grains of salt.

Other studies have dramatically different findings, with the Extracorporeal Life Support Organization registry suggesting a 90-day in-hospital mortality rate of 38% with VV-ECMO for patients with COVID-19 acute respiratory distress syndrome (ARDS), and another suggesting just 56.8% survival to discharge, with the majority of patients going to other hospitals, long-term care, or rehabilitation facilities -- poorer results than those seen for non-COVID ARDS.

"These discordant outcomes likely reflect differences in patient selection, patient management, and a willingness to continue support in the setting of single-organ dysfunction," Smith's group acknowledged.

Many other centers used VV-ECMO for anyone who was decompensating as a lifesaving measure despite knowing many wouldn't survive, Mitzman pointed out.

But rather than a last-ditch strategy, NYU Langone offered VV-ECMO to "patients that the team believed had a reasonable chance for survival," with the aim to initiate it early in the clinical course and limit exposure to the effects of high airway pressure, respiratory rates, and levels of oxygenation.

Standardized, aggressive management also included "not deviating from lung-protective ventilation strategies, nearly universal early tracheostomy and frequent bronchoscopy, treatment of coinfection, and standardization of an anticoagulation regimen," Smith's group added.

During the time of uncertainty and limited resources in the initial wave of the pandemic, NYU Langone evaluated 80 patients for VV-ECMO and treated 30 with it out of the 415 COVID-19 patients admitted to the ICU. These patients were hospitalized for a median of 45 days and supported on VV-ECMO for a median of 19 days.

"We were struck by the number of young, otherwise healthy patients who were dying from the disease. Because of this, we felt patient selection was very important," Smith said in a statement.

Mitzman characterized the cohort as the "youngest and the healthiest of the sickest," with only single organ failure. Median age was 42, 87% were men, 53% were obese, and 60% were Hispanic.

At 1 year, there were some signs of residual pulmonary damage in that the majority of patients had a diffusing capacity for carbon monoxide level at the lower end of normal, but the researchers noted that its clinical significance is unknown.

"Although available data thus far are limited to approximately one-half of the cohort, a median 6-minute walk test of 350 m is encouraging," Smith's group wrote.

That was the Achilles heel of the data, Mitzman suggested. Only 44% of survivors to hospital discharge had pulmonary function testing at 1 year, and 59% had a 6-minute walk test.

Often, the people who are willing to come back and get tested are the ones who are doing well, he told MedPage Today. "Most likely, the patients who weren't tested, that they don't have the results on, are patients that were still having symptoms and may have had decreased pulmonary function. So this leads to significant information bias."

The observational study design limited what lessons can be drawn as well, he said.

"We can't say that early ECMO in these extremely ill patients will lead to improved survival or no long-term pulmonary function issues based on these data," he noted. "There's too much selection bias."

The researchers, though acknowledging those limitations, emphasized the importance of permitting sufficient time to recover on ECMO for COVID-19 ARDS rather than moving on to lung transplantation after as little as 1 or 2 months. Even if patients do go on to need donor lungs, they wrote, "a period of rehabilitation, even if these patients are oxygen dependent, may result in more rapid recovery from their eventual transplantation."

Comment