The Doctors Who Bet Their Patients’ Lives on COVID-19 Test Results

Those waiting for an organ transplant are at risk of contracting the coronavirus whether they choose to avoid the health-care system or to interact with it.

Chest X-ray with lungs highlighted
Adam Maida / The Atlantic

When the third coronavirus surge hit the U.S. last fall, the midwestern states were among the worst affected. Thousands of people in the region were being hospitalized with the virus every day. It was at this inauspicious time that a team of transplant doctors at University Hospital in Ann Arbor, Michigan, received a pair of healthy-seeming lungs. According to a published case report, the donor had been in an automobile accident, and died from her injuries a few days later. She’d shown no signs of being sick, according to her family, nor had she been knowingly exposed to anyone with COVID-19. A radiologist did find an abnormality in her right lung but chalked it up to damage from the accident. Meanwhile, a nasal swab, taken at the hospital, confirmed her infection status: She was negative.

The patient for whom those lungs were meant to be lifesaving—a woman with chronic obstructive lung disease—also tested negative for COVID-19, in a nasal swab taken 12 hours before her surgery. But three days later, the recipient was in severe distress: She was feverish, with plummeting blood pressure, and she experienced such difficulty breathing that she had to be placed on a ventilator. Now she tested positive for the coronavirus. (One of the transplant surgeons, too, would end up sick.)

In the weeks that followed, the transplant patient received the best available COVID-19 treatments, including remdesivir and convalescent plasma, but doctors couldn’t save her. Two months after the procedure, she was dead. A reexamination of respiratory fluid taken from her donor before surgery revealed the source of the infection: The transplanted lungs that doctors sewed inside her body had been teeming with the coronavirus.

I’m a physician who specializes in diagnostics, so one quirk of my pandemic experience has been getting lots of text messages from my friends about the polymerase chain reaction. PCR is used in laboratories to identify everything from genetic conditions to infections to cancers, though you probably know it as the “gold standard” method for detecting SARS-CoV-2, the virus that causes COVID-19. The friends who texted me had gotten COVID-19 tests, and they wanted help interpreting their results.

A COVID-19 test result seems straightforward on its face: You’re either positive or negative. But questions often follow: What if you feel a little sick, but your test has come up empty—can you risk going back to work? Or what if your test is positive, but you feel completely fine—should you repeat it later to confirm that you’ve recovered? How much later? In the early months of the pandemic, I muddled through giving advice to friends based on what was already known about the technology, and on the preliminary data coming out of China. Probably not, I said. Maybe so. A week or two? I couldn’t say very much with certainty.

A year later, my colleagues and I have more and better facts to help us through this diagnostic slop. Now we know that a positive result on a PCR test won’t tell you whether you’re currently contagious, but it can say—with 99.9 percent accuracy—that you’ve been infected with SARS-CoV-2. As for false negatives, larger analyses suggest that about one in eight infections could be missed.

There’s one specific branch of medicine where even these modest risks of error simply cannot be abided. For the more than 107,000 Americans who are now waiting for an organ transplant—and for those who have already received an organ—the stakes of COVID-19 testing are amplified many times over. It’s easy to understand how a missed infection in a donor could lead to deadly complications for the transplantee, as in the tragedy last fall. But a false-positive result—a COVID-19 case that isn’t real or is long-recovered—may be fatal too, when it delays or prevents an organ from reaching a desperate patient. It’s hard enough for frontline doctors to interpret a surprising test result. For those who work in transplant medicine, decisions made under this uncertainty could be irreversible.

“It’s my worst nightmare,” Joshua Lieberman, a pathologist at the University of Washington who works on transplantation testing, said when I asked about the case in Michigan. He was particularly struck by the extent of infection found in the donated lungs. “There’s not a little bit of COVID in there. It is rip-roaringly positive,” he said of the PCR results—“like, a million times more virus” than he usually sees.

How could that infection have been missed? Even at the start of the pandemic, we knew that patients could be admitted to a hospital with severe breathing problems but get a negative COVID-19 test result. One very early study, done in China, found that sick patients might test negative in samples taken from the nose, but positive in fluid from the lungs; it’s since been confirmed that a lung sample can catch about 13 percent more infections than a regular, nasopharyngeal swab. That’s why the American Society of Transplantation recommends this form of testing for every lung donation.

Still, organ-procurement organizations have held off on mandating this. Not every lab can process fluids from the lung, Lieberman told me, so adding this one requirement might end up reducing patients’ access to a scarce resource. Lifesaving surgeries could be delayed.

The problem is that even a very modest COVID-19 infection in a patient who has received a new organ has profound implications, Ajit Limaye, an infectious-disease physician at the University of Washington, told me. Patients who catch the coronavirus around the time of any sort of surgery are at a tripled risk of death. Those who have received new organs are still more susceptible, on account of their being sick enough to need a new organ, and their being on powerful immunosuppressive drugs to prevent rejection of that organ. One review found that among transplant recipients who became infected with the coronavirus, 81 percent needed to be hospitalized. (That hospitalization rate for the general U.S. population is estimated to be about 5 percent.) Even the usual rule of thumb for how long someone might remain actively infected are thrown out the window for transplant patients. While a case of COVID-19 typically lasts about two weeks, live virus has been recovered from immunosuppressed people more than two months after their initial infection.

But being overcautious about potentially infected donors can be deadly too. There is no safe option for patients in the medically fragile state of organ failure: A recent study found that people waiting for a kidney during the pandemic have been at a 37 percent greater risk of dying than people who were on the list before. Because the queue for kidneys is so long—about 90,000 people—this finding could mean “a substantial number of additional deaths,” the authors wrote. This is the catch-22 of COVID-19. The sickest people, such as those in organ failure, are at risk whether they choose to avoid the health-care system or to interact with it.

Medical decisions involving organ donation are made all the more challenging by the time constraints involved. In the hours it takes to sort out an intended transplant recipient’s true infection status after a first, positive result—by repeating the test on a different PCR machine, for example, or scrutinizing how much viral material was detected—an organ may need to be diverted to someone else or discarded entirely. Limaye knows of transplants that have been canceled because doctors didn’t have a quick (enough) way to determine whether a patient’s infection as picked up by PCR actually posed a risk to them or others. Could the test have spotted a case of COVID-19 that was already resolved? Were the doctors willing to bet someone’s life on that presumption?

Laboratory experts have developed some helpful ways to pick up false positives and false negatives on COVID-19 tests, from transplant patients and others, too. One thing they look for is an incongruent result. PCR machines check a sample for matches to multiple elements of the virus’s genetic code. When only one of those elements is detected, doctors might repeat the test to make sure it’s accurate. A microbiologist named April Abbott has also pointed out that, in rare cases, a sample’s viral load is so high as to be literally off the charts—and thus invisible to laboratory software. That problem can be solved, she said, by looking at the analyzer’s raw data, not just its automated result.

Beyond a few simple improvements, though, there are no easy answers for doctors overseeing transplants. In these situations, a test result may serve as the basis for a life-or-death decision. Taking a careful medical history can help mitigate the risk of misdiagnosis, when combined with a physical examination and the results of other laboratory or imaging studies that have already been performed. Doctors then pull together all of this information to estimate a patient’s “pre-test probability” of infection—how likely they might have been to have COVID-19 before their swabs were sent off for testing. A probability isn’t a certainty, however. The Michigan doctors had been using just this logic when they decided to forgo any further testing on the lung donor: Because her initial swab was negative and she lacked any known symptoms or exposures, they determined that the risk of missing an actual infection was very low.

Organs from deceased donors, which make up the majority of transplants in the United States, present a unique challenge for this careful clinical interpretation. Doctors cannot take a medical history from a patient who suffered brain death after a car accident or drug overdose. (In the case of the Michigan donor, the medical history was provided by the family.) Many of these potential donors are young and relatively healthy, so if they had been sick with COVID-19, they’re more likely to have had a mild case—and therefore to have been unaware of it. “They may have never been tested or had any symptoms,” Limaye said. “So we’re left with limited information.”

Meanwhile, any positive test result from a deceased donor will probably mean their organs are removed from the system for good. Limaye worries that this “COVID Zero” approach may not be the right one for transplant medicine. There are circumstances, he argues, in which it would be worth the risk to allow an organ donation even from someone with a known infection. (The rules are more relaxed for living donors, who are generally allowed to share their organs three weeks after initial signs of COVID-19, even if follow-up testing still registers as positive.) For example, a patient who may not survive without a rapid transplant could benefit from receiving an organ from a deceased donor who had experienced only a mild or asymptomatic infection. One published case series looked at transplants drawn from six deceased donors who had tested positive for COVID-19 at some point before their death; none ended up transmitting the virus.

The complexity of these clinical judgments is likely to persist, even as more Americans are vaccinated. More than half of all U.S. adults have now received at least one dose of a COVID-19 vaccine, but transplant patients are in a special group, left to navigate the pandemic with added uncertainty. Studies of both the Pfizer and Moderna vaccines excluded people who are on immunosuppressive medications, so we don’t yet know how effective the shots might be at preventing illness in an organ recipient. Johnson & Johnson did include some transplant recipients in its clinical trial, but only a handful. One study looked at 658 organ-transplant recipients who had been fully vaccinated with either the Pfizer or Moderna vaccine, and found that only about half of them produced detectable levels of the relevant antibodies—in contrast to nearly 100 percent of healthy people.

That’s not the only “data void,” as Limaye describes it, for transplant doctors. They don’t know exactly when it’s safe to transplant an organ after a donor’s positive COVID-19 test, and they don’t know whether some organs from infected donors might be safer to transplant than others (a lung from someone with a respiratory virus might be more dangerous, for example, than a kidney or a liver). Until those gaps in knowledge can be filled with rigorous, unbiased research, doctors can only keep a broad perspective on the stakes involved. What would be the consequences of approaching a test result with too much caution—or with too much chutzpah? Deciding whether it’s more important to guard against false negatives or false positives, as one transplant doctor put it, may be a matter of deciding “what scares you more.”

A COVID-19 diagnosis is powerful, and the downstream consequences can’t always be predicted. Diagnostic labels, like medical interventions, may be lifesaving or life-threatening in themselves. It’s likely that more intense screening of transplant patients has prevented other tragedies like the Michigan case, but it may also have cost some people a new organ. Practicing medicine means imperfect answers and inevitable trade-offs. “There’s nothing sacred about COVID tests,” Limaye said. “We’re learning that they have challenges of interpretation, like virtually every single test we do.”