During a surge of coronavirus cases at Houston Methodist Hospital last summer, a patient in his 40s on a ventilator was declining. There was one more option, a last-resort treatment that can mechanically substitute for badly damaged lungs.

But that day, the slots designated for the intensive treatment, called ECMO, were filled. One patient, a man a decade older, had been receiving the therapy for over a month. Doctors had concluded he had almost no chance of recovery, and had recommended stopping the treatment several times, but his relatives were not ready to let him go.

“We have to push some more,” said Dr. Sarah Beshay, a critical care physician, because the younger patient “needs a chance too.”

That afternoon, she called the older man’s daughter, who had not been allowed to visit because of COVID restrictions. Explaining that the therapy was in scarce supply, the physician said, “It’s a matter of using the available resources in the wisest way possible.”

The daughter interrupted, bluntly asking if the doctors were trying to remove the equipment from her father to give it to someone else. Beshay said no, adding that it was a physician’s duty to inform a family that persisting with treatment was “not the right thing from a medical perspective” when chances of recovery were minimal.

The next day, the family agreed to withdraw ECMO and he died. A day later, two patients were successfully taken off the treatment after improving, and others started on it, including the man in his 40s. A month later, however, doctors were having a difficult conversation with his family.


Throughout the pandemic, such scenes have played out across the country as American doctors found themselves in the unfamiliar position of overtly rationing a treatment. But it was not ventilators, as initially feared: Concerted action largely headed off those shortages. Instead, it was the limited availability of ECMO — which requires expensive equipment similar in concept to a heart-lung machine and specially trained staff who can provide constant monitoring and one-on-one nursing — that forced stark choices among patients.

More on the COVID-19 pandemic

Doctors tried to select individuals most likely to benefit. But dozens of interviews with medical staff and patients across the country, and reporting inside five hospitals that provide ECMO, revealed that in the absence of regional sharing systems to ensure fairness and match resources to needs, hospitals and clinicians were left to apply differing criteria, with insurance coverage, geography and even personal appeals having an influence.

“It’s unsettling to have to make those kinds of decisions,” said Dr. Ryan Barbaro, a critical care physician in Michigan and head of an international registry of COVID-19 patients who have received ECMO — short for extracorporeal membrane oxygenation — about half of whom survived hospitalization.

“Patients died because they could not get ECMO,” said Dr. Lena M. Napolitano, co-director of the Surgical Critical Care Unit at the University of Michigan. This spring, she was overwhelmed with requests to accept patients considered good candidates for the therapy. “We could not accommodate all of them,” she said.

And despite the progress the United States has made against the virus, some doctors are still having to ration ECMO, which is offered in less than 10% of hospitals.


“It’s something we’re balancing every day,” said Dr. Erik Eddie Suarez, a cardiovascular surgeon at Houston Methodist. If the hospital accepts too many COVID patients for ECMO, he said, “we can’t do cardiac surgery,” because some of those patients also need the treatment.

Of the more than 185 million known coronavirus cases worldwide since December 2019, close to 8,000 patients have received ECMO to date, including nearly 5,000 in North America, according to a registry maintained by the Extracorporeal Life Support Organization.

Among them were a family doctor and a police sergeant nearing death in Southern California whose cases demonstrate both the promise and the clinical and ethical challenges of the therapy.

Dr. David Gutierrez, 62, cared for patients with coronavirus in a high desert town northeast of Los Angeles before catching it last winter. But the hospital where he was gravely ill did not offer ECMO, and others nearby that did were full or would not take him. “My father had no options,” said Dr. David Gutierrez Jr.

The elder Gutierrez was beyond the age cutoff established by Providence Saint John’s Health Center in Santa Monica, California, during the coronavirus surge and had underlying health conditions that decrease ECMO’s chances of success. Physicians there accepted him in January anyway, partly because of the risks he had taken caring for patients, said Dr. Terese Hammond, head of the intensive care unit. During the peak, the hospital had 11 COVID patients on ECMO at one time; as of Sunday, it had three.

During surges there and elsewhere, securing a precious ECMO slot often required extraordinary advocacy by a patient’s family, colleagues or medical providers.


That was the case for Los Angeles police Sgt. Anthony Ray White, an athletic, 54-year-old father of two with Type 2 diabetes whose department sent him for coronavirus testing after a potential exposure on the job in late December. When he fell ill, he was treated first at a Kaiser Permanente hospital that did not offer ECMO. The medical team there told his family that he would die, that it was time to withdraw care and say goodbye.

His wife and sister refused to accept the prognosis. The medical director of the Los Angeles Police Department intervened, persuading the hospital to allow White to be transferred to Saint John’s. “They’re my family so I advocate for them,” Dr. Kenji Inaba said of the police department. He said he had pushed to get ECMO for several other officers who almost certainly would have died without it.

“A person shouldn’t have to be a police officer or have connections to get health care,” said Twila White, the sergeant’s sister.

Finding a ‘Goldilocks’ moment

After the coronavirus struck China, some doctors there used ECMO to treat COVID-19 patients, but they reported poor outcomes — 80% of patients in one Hubei, China, study died. When the disease exploded next in Italy, doctors were overwhelmed and did not try it much. That March, the Swiss Academy of Medical Sciences recommended against giving the treatment to COVID patients.

ECMO involves a bedside surgery to connect major blood vessels with equipment that adds oxygen and removes carbon dioxide from the blood before pumping it back to the patient, allowing the lungs or heart to rest. One to two highly trained nurses care for each patient, with respiratory therapists and often with technicians known as ECMO specialists or perfusionists.

When cases began rising in New York last March, ECMO teams were “flying blind,” said Dr. Mangala Narasimhan, a director of critical care services at Northwell Health, New York’s largest medical system. Unsure of whether to offer ECMO, staff members debated potential risks and benefits.


While a course of ECMO often lasts four or five days for respiratory failure, doctors learned that COVID patients could require weeks. “You’ve got to figure out, do they really need it and is it really enough,” Narasimhan said. Out of roughly 14,000 COVID patients treated in the hospital system during the initial surge — close to 2,500 in intensive care — only 23 were put on ECMO, with about 60% surviving, she said.

One day last April at Long Island Jewish, a flagship Northwell hospital, Narasimhan was called multiple times to consider potential ECMO patients. That week, roughly 900 suspected or confirmed coronavirus cases packed a facility whose usual bed capacity was 583. One bad day, 84 patients died.

Across the hospital system, seven patients were on ECMO for lung failure; normally there would be one or two. Narasimhan went to evaluate a 60-year-old with diabetes and heart disease who had COVID and was faring poorly. The physician turned the patient down for ECMO, given the age and existing health conditions.

Narasimhan then discussed a 20-year-old at a hospital roughly an hour away. “The patient’s already dying,” she told a colleague. Team members could go get the patient, but “they don’t think they can get up there fast enough.”

Putting a critically ill patient on ECMO requires finding what Dr. Subhasis Chatterjee of Baylor St. Luke’s Medical Center in Houston called the “Goldilocks” moment — not too early, when less intense therapies may still work, but also not too late, when too much damage has occurred.

As coronavirus patients flooded Houston Methodist Hospital last summer, officials set a cap of eight COVID patients on the therapy at any time, even though there were additional ECMO devices — in part to reserve capacity for heart surgery patients, and because nurses reported that they could not safely care for more. But the prospect of watching good candidates for ECMO die was excruciating. In mid-July, Suarez, the cardiovascular surgeon, started a patient on the treatment despite having been told not to because the ceiling had been reached. “The man was dying in front of me, and we had the machine,” he said. The patient survived and made it home.


As cases continued to rise, the hospital created a daily process to triage ECMO, which included input from ethicists. Doctors specializing in end-of-life care worked with family members to help prepare them for the possibility that their loved ones would not recover, and they were allowed to visit before a dying patient was taken off ECMO.

Still, the dilemmas have persisted. Houston Methodist, which has treated 90 COVID patients with ECMO, turned down roughly 120 requests for it just this year, mostly for lack of capacity, according to the head of critical care, Dr. Faisal Masud. He said he expected demand to remain high because of unvaccinated residents and the treatment’s broader utility for lung failure.

Dr. Jayna Gardner-Gray, a critical care and emergency physician at Henry Ford Health System in Detroit, said during a surge this spring she kept asking herself how long to keep patients on ECMO when it appeared, but was not certain, that they would never recover. “If no one else was waiting, would I let them go?” she said.

Dr. Antone Tatooles works at two Chicago-area hospitals that initially had good success with ECMO. But when one of them took on more COVID patients, survival rates fell. “We got overwhelmed,” he said. “We can apply technology, but we need appropriate human resources.”

Outcomes vary widely among hospitals. Overall, however, survival has decreased over time, including at major U.S. and European hospitals. From January to May of 2020, according to the international registry, less than 40% of COVID patients died in the first 90 days after ECMO was started. But in the months after that, more than half died. “The patients seem to be doing markedly worse,” Barbaro said.

He and his colleagues are analyzing whether that relates to factors like new virus variants, less experienced centers providing care or changes in the treatments patients receive before ECMO.


Who can pay, and who can’t

ECMO is offered in few community hospitals, where most Americans get care. Saint John’s, the Santa Monica facility where the doctor and police sergeant received the treatment, is an exception.

It started an ECMO program about a year before COVID-19 emerged. The 266-bed hospital has provided the therapy to 52 COVID patients during the pandemic, about the same as the entire Northwell health system in New York, which has more than 6,000 hospital and long-term-care beds.

The Saint John’s charitable foundation, supported by the area’s wealthy donor base, helped fund the ECMO program and its expansion. The hospital accepted some uninsured COVID patients for ECMO, whereas elsewhere these patients were often turned down despite a federal program that reimburses hospitals for their care.

“There are just so many inequities,” said Hammond, Saint John’s ICU director. And for every COVID patient who survived with ECMO, there are “probably three, four, five people that die on the waiting list.”

She and other doctors said the pandemic highlighted the need for ECMO to be more widely available and less resource intensive. Until then, “we really need to have a system for sharing,” she said. Allocation systems do exist for transplant organs and trauma care.

Getting patients moved to a hospital with ECMO often depends on relationships between doctors and having a case manager “who really knows how to push,” said Dr. Michael Katz, a critical care specialist at St. Jude Medical Center in Fullerton, California, who has transferred patients elsewhere for ECMO. In multiple cases, he said, by the time a hospital had financially evaluated the patient’s insurance status, it was too late.


By contrast, Minnesota’s ECMO centers formed a consortium and issued standard eligibility criteria to help ensure that every patient had “the same shot” at getting the therapy, said Dr. Matthew Prekker, the ECMO medical director at Hennepin County Medical Center. “There weren’t any double standards. No one had to go ECMO shopping.”

Centralized ECMO triage systems also exist in Britain and the Paris metropolitan region.

Throughout the world, the main considerations for selecting patients have been medical ones, and the organization that maintains the ECMO registry offers guidelines. During a surge in cases, individual institutions often tightened the criteria.

That mostly involved lowering age limits, as Saint John’s did, moving its cap from 70 to 60.

There were some exceptions, like 62-year-old Gutierrez, who loved Netflix and Korean dramas and was soon to become a grandfather.

Gutierrez had a rocky course on ECMO. Fluid collected around his heart. He bled easily, developed other infections and required kidney dialysis.


In February, he improved enough for the medical team to stop ECMO. Still, he faded in and out of consciousness and continued to require a ventilator.

His wife and his two adult children visited, and other relatives joined a daily prayer call. Over several months, his lungs began to heal. He spoke between huffs, closing his eyes with the effort. Without ECMO, he said, he would probably be dead.

In June, medical staff at his rehabilitation facility clapped as he was discharged home in time for Father’s Day. He remains weak, but aims to be treating patients again by January.

White improved after transferring to Saint John’s for ECMO. Doctors woke him up, and he engaged in video calls with his school-age children and his wife, Tawnya White.

But his lungs did not recover, and in late February, he was transferred to UCLA for a transplant evaluation. To qualify, he had to get strong enough to walk, and test negative for the coronavirus and other infections. He wrote on a white board that he was hoping to “get well for retirement.”

Back at their family home in Eastvale, about 50 miles from the city, Anthony White’s 11-year-old son recalled his father teaching him to play chess. His 15-year-old daughter spoke wistfully of going out for fast food with him after soccer practice. He and her mother would sing as they cooked together, near a sign that read, “This kitchen is for dancing.”


“He is a beautiful person with a beautiful heart,” his wife said.

Soon he could sit in a chair, and in March, he stood for the first time in months. He even took a few steps, and doctors hoped that his lungs might yet heal.

But setbacks chased every milestone. A chamber of his heart malfunctioned. His oxygen levels dipped. He developed an unusual fungal infection.

The weeks passed in a painful limbo for Tawnya White, who was not allowed to visit. “I go to bed thinking about him, I wake up thinking about him,” she said.

At last, in April, the hospital loosened its no-visitor policy. Her husband took her hand, and she read his lips as he tried to speak: “How do I get stronger?”

Two days later, Anthony White had his second consecutive negative coronavirus test. He had finally cleared the infection. But two days after that, his 100th day of hospitalization, doctors told Tawnya White her husband was dying. He had developed an aggressive bacterial pneumonia. Within days, he was gone.

His wife takes comfort that he was given his best chance at survival. “I’m still at peace that everything possible was done for him.” she said. But she feels it is unfair that was not the case for other patients. “Everyone should have access to everything that Anthony had,” she said.

This article originally appeared in The New York Times.