A decades-old lab procedure could help the United States claw its way out of yet another crippling coronavirus testing shortage, but the nation is only beginning to ramp it up now.
The method, called “pooled testing,” allows labs to test combinations of multiple swab samples in batches. If the virus turns out to be present in a batch, then each of its samples gets retested to determine which are positive. But if a batch is negative, then all of its samples are considered virus-free and don’t need to be tested individually, saving both time and scarce supplies.
In the US, which has more than 4 million confirmed cases and many more unconfirmed, public and private laboratories are swamped. Turnaround times of as long as 19 days for test results are making it all but impossible to contain the virus.
Proponents say that pooled testing, which has already been deployed in China, Israel, and Germany, could be one of many innovations that would enable the US to scale up testing. And it might be particularly useful in settings where large numbers of people require routine tests, such as nursing homes and colleges welcoming back students in person in the fall.
But to the frustration of some public health experts, the FDA only just issued its first emergency authorization that allows a company to do this, months into the pandemic. In mid-July, Quest Diagnostics got cleared to test pools of suspected coronavirus samples. Days later, LabCorp was authorized to pool samples from people without symptoms. Meanwhile, a handful of other labs have been allowed to do pooling while they apply for authorization.
“Pooling just becomes a way to stretch capacity,” Carl Bergstrom, an epidemiologist at the University of Washington, told BuzzFeed News. “Why hasn’t there been more of a push to implement pooled testing at a commercial scale?”
The FDA, which is in charge of ensuring the accuracy of diagnostic tests, points out that pooling has limitations. It’s less effective in places where infections are widespread, for example. And the more samples that are pooled together, the greater the chances are that a positive result will be missed.
Those weaknesses aside, experts say that pooling could still help the country significantly ramp up testing at a time when it’s desperately needed. Admiral Brett Giroir, the Department of Health and Human Services official overseeing testing, has said he wants the US to be conducting 100 million monthly tests by September, and that pooling could help it get there.
“Covid-19 is ravaging the US,” Manisha Juthani, an associate professor of medicine and epidemiology at Yale School of Medicine, said by email. “We need to find as many different strategies that can quickly identify cases and quarantine those people to prevent spread of disease. Pool testing is worth a shot.”
Pooled testing could help us ramp up testing — but it can’t be used everywhere.
The idea of pooled testing dates back to World War II, when it was suggested as a way to test soldiers for syphilis. Since then, it’s been used to screen for a variety of sexually transmitted and infectious diseases, such as HIV, hepatitis B and C, West Nile virus, and the avian flu H5N1.
Pooling can detect the coronavirus, too, though only in places where the virus is at low levels.
Juthani explained it this way: Suppose the virus is present in 5% of a community, meaning five out of 100 samples are true positives. You could run 100 individual tests to find the five positives. It would be more efficient, however, to test them in batches of, say, 20 pools.
If the five positive cases were scattered across different pools, then five groups would come back positive. Each sample in those pools would then need to be retested in order to surface the positive ones. The number of tests run in this scenario would total 45, a significant saving of supplies and time.
But this math breaks down when the virus is rampant. Suppose a community’s transmission level is 30%, meaning 30 out of 100 samples are expected to be true positives. If you tested this group of 100 tests in batches of 20, there’s a good chance that one positive sample would be in every batch. Then you’d end up having to retest all 100 samples anyway, on top of testing the 20 pools. That, Juthani said, would be “a waste of time.”
“If we had suppressed the virus from the start, pooled testing would have been appropriate everywhere in the US right now,” Juthani wrote. “Given that we didn’t do that, pooled testing is only appropriate to use in certain locations.”
So while pooled testing wouldn’t be helpful right now in states like California, Texas, and Florida, it could work in, say, the Northeast, where transmission is much lower than it was this spring. But decreasing the burden in regions or states with lower infection rates could still help increase testing capacity overall, experts say.
The American Clinical Laboratory Association, which represents private labs like Quest and LabCorp, said that its members have quadrupled their capacity in the last three months to more than 400,000 tests a day. “We’re focused on continuing to build that capacity, and pooling is one facet of that effort,” said Julie Khani, the group’s president, through a spokesperson.
Quest presented the FDA with data showing that its pooled testing wouldn’t miss any infections as long as the virus’s prevalence in a population was below 10%, according to a spokesperson. The lab-testing giant is currently pooling in its labs in Chantilly, Virginia, and Marlborough, Massachusetts, and plans to roll it out elsewhere on an unspecified timeline.
The FDA authorized Quest to pool four samples from people suspected of having SARS-CoV-2, and LabCorp to pool five samples from people without symptoms. In theory, testing could become even faster if the pools were even bigger.
One hurdle is that if more samples get lumped together, the risk that the virus will be diluted down to undetectable levels becomes greater. That’s even more likely if a sample has a low level of the virus to begin with.
Some scientists believe there can be workarounds. One paper, published in June, calculated that pools could have as many as 25 samples and still detect about 80% of the positive cases. This would work as long as the virus was rare enough, according to the study’s lead author, Christopher Pilcher, who helped pioneer the use of pooling for HIV detection in North Carolina.
Given the significant logistical work it takes to combine, separate, and retest samples, he thinks that special pooled testing facilities should be set up to shoulder much of the coronavirus testing burden.
“You could get next-day turnaround on the specimens, and remove half the specimens from the testing queue and current labs that are being crushed right now,” said Pilcher, a professor emeritus of medicine at UC San Francisco. “That’s the promise of pooling right there.”
But the FDA has not been as quick to embrace pooling as some experts would like.
As the pandemic spread across the US in March, Nebraska’s public health laboratory, like so many others, was running out of supplies. Director Peter Iwen thought pooling might be able to help, so he went ahead and ran preliminary tests in groups of five — even though he didn’t technically have permission to do so. It worked.
Iwen was using a diagnostic test made by the CDC, and if he wanted to make any changes to how it was used and report those results to patients, he was supposed to get the FDA’s authorization first. When he asked both agencies for permission to use the test on pooled samples, the CDC gave a nondefinitive response, then went silent. The FDA didn’t reply right away.
With reagents dwindling, he made a last-ditch effort and sent his data to Nebraska’s governor, who in a state of emergency has the ability to supersede federal law. The governor gave his approval. A few days later, the FDA signed off too.
“What it did is it allowed us to keep a lot of COVID-19 testing going in Nebraska,” said Iwen, who estimated that pooling saved more than 50% in supplies and 25% in staff time. “Otherwise, we would have had to shut down.”
Nebraska’s transmission rate was low at the time. The outbreak there has since picked up significantly, due to an explosion of cases in meatpacking plants. Iwen said the lab stopped pooling around May 1, once the positivity rate hit about 10% and “it just wasn’t worth it for us to do it anymore.”
Iwen and other public health experts said they are bewildered by the FDA’s oversight of coronavirus tests. The agency was widely criticized for letting dozens of unverified antibody, or serology, tests flood the US this spring, but, in Iwen’s view, has been overly cautious with allowing pooling.
In mid-June, the FDA released guidance for the multiple studies that laboratories and companies need to perform before they can run diagnostic tests on pooled samples. Iwen said that lab directors around the country have told him they want to start pooling, but don’t have enough reagents to do the studies required — which, he observed, is “kind of a catch-22.”
Iwen wants the FDA to loosen up on its requirements, especially for public labs and smaller labs at hospitals. “They don’t want garbage reported and I understand that,” he said. But “I’m having a hard time figuring out where they’re coming from.”
FDA spokesperson Emma Spaulding said that the agency “has always been open to working with diagnostic developers on novel testing ideas, like specimen pooling.” She added, “We continue to proactively work with developers to help facilitate new approaches and meet the testing needs of Americans.”
The FDA is in a tough position, said Kelly Wroblewski, infectious disease director at the Association of Public Health Laboratories.
“They moved really quickly on things earlier in the response, like serology testing, that they then had to spend a significant amount of time and effort walking back,” she said. “I don’t blame them for proceeding with some caution now.”
But to Iwen, pooling is one badly needed means of chipping away at a caseload that grows higher every day.
“We’re running out of reagents again. It seems like everybody in the world wants to get tested,” he said. “Something has to be done.” ●
Correction: Peter Iwen, director of Nebraska's public health laboratory, ran tests on pools of five samples. An earlier version of this story misstated the number of samples.
This story has been updated to include further details from Peter Iwen about the regulations over modifications to diagnostic tests.