NEW YORK — Pool testing for SARS-CoV-2 has been touted as a solution for screening during the COVID-19 pandemic, but while new ways to carry out such testing abound, questions remain about sensitivity issues and the regulation of such tests.
Earlier this week, the US Food and Drug Administration provided guidance for pool testing of samples to enable broader screening of asymptomatic individuals who might be infected with SARS-CoV-2, the virus that causes COVID-19. In its June 16 statement, the US regulatory agency outlined steps for such group tests to receive authorization for clinical use.
Pool testing has been an established approach in diagnostics for decades, but it has taken on new significance during the COVID-19 pandemic as testing resources such as reagents have been constrained by unprecedented demand and labs have struggled to scale up testing.
To overcome these challenges, some labs have decided to combine samples into batches or pools and then test them with a single test. If the result is negative, then all of the samples in the pool are presumed to be negative for the virus. But if the result comes back positive, it means at least one sample is positive and they all require retesting.
In its updated template for molecular tests, the FDA provided recommendations for offering pool testing for SARS-CoV-2, including specifications for using an assay that has or has not yet received Emergency Use Authorization. These include metrics for sensitivity and validation studies, as well as guidance on generating data at geographically diverse sites, for example.
For researchers who have been honing new pool testing methods for SARS-CoV-2, the fresh guidance from the FDA was very welcome.
"I think it is a good move on the FDA's part because we need to screen a lot of asymptomatic people such students, workers, and military personnel, so pooling is really important for its cost saving and faster turnaround time," said Baha Abdalhamid, an associate professor at the University of Nebraska Medical Center.
"They also addressed assay sensitivity with pooling, which is a significant issue," Abdalhamid said. "That is why labs need to perform a limit of detection [study] for the sensitivity of the assay."
Abdalhamid co-authored one of many recent papers to address implementing pool testing for SARS-CoV-2 since the pandemic began late last year. In April, he and colleagues at UNMC published their approach in an MedRxiv preprint, which has since been published in the American Journal of Clinical Pathology.
In the paper, they set out to determine the optimal parameters for group testing of samples using the Centers for Disease Control and Prevention's SARS-CoV-2 RT-PCR assay. Using a web-based algorithm developed at the University of Nebraska, they determined the most efficient pool size to be five specimens and tested the approach on 25 pools, each containing one positive and four negative samples, as well as another group of unselected 60 specimens grouped in 12 pools. All of the 25 pools were determined to be positive using the approach, while the analysis of the additional 12 pools determined that two were positive. In total, the researchers ran 22 tests versus the 60 tests they would have needed to run for each individual specimen.
"The pooling test is a straightforward, cost-effective, and time-saving approach," said Abdalhamid. "It conserves reagents and material and expands the test for more people."
A number of factors need to be considered when setting up pool testing, he said. One is the prevalence of COVID-19 in the area where testing is being implemented. "As a principle, the lower the prevalence, the more specimens you can pool and the more effective the approach is," noted Abdalhamid. "At a prevalence rate of 5 percent, we pooled five specimens," he said. "At a prevalence rate of 1 percent, you can pool 11 specimens."
He also stressed the importance of sensitivity and specificity, as well as algorithm selection. Sensitivity is particularly critical because when a sample is diluted during the pooling process, a less sensitive assay might fail to yield a positive result.
But regulatory approval is also a hurdle, he noted, as all COVID-19 testing in the US is being controlled through the FDA's EUA process, which he called "time consuming." In other regions, such as Europe, diagnostics for SARS-CoV-2 have largely been overseen at the national level, meaning that some labs have had more leeway to adopt and experiment with the testing approach.
Yet just because one lab implements pool testing, it doesn't mean that another lab can copy it, as pool sizes have to be calculated based on disease prevalence within the region, Abdalhamid noted. This means that every lab that implements pool testing has to run its own validation study.
Looking for the perfect approach
According to Jens Eberhardt, a researcher at the Max Planck Institute for Mathematics in Bonn, Germany, the first literature on pool testing dates back to World War II, when Robert Dorfman developed the method, also called Dorfman testing, to screen American soldiers for syphilis.
Dorfman initially proposed a two-stage pool testing method, where samples are tested in pools in a first stage, followed by a second stage of individual testing. Yet Eberhardt and colleagues proposed a multi-stage approach for pool testing of SARS-Cov-2 in a MedRxiv preprint in April, which has since been published in the Journal of Clinical Virology.
In the paper, the authors discussed the design of different multi-stage testing approaches and compared their efficiency at different prevalence rates using computer simulations. They determined that three-stage testing schemes with pool sizes of up to 16 samples could test many more samples than a two-stage approach with the same resources and proposed an adaptive approach, where the optimal testing scheme is selected based on the expected prevalence rate.
"Our paper investigates more sophisticated pooling methods, which can have more than two stages," Eberhardt said, where the samples are divided into smaller subpools. "For example, one could pool 16 samples together in a first stage, then proceed to four pools of four samples in a second stage, and then test individually in a third stage," he said. "This raises the efficiency of pool testing significantly in comparison to Dorfmans's classical approach."
While Eberhardt's team sought to improve pool testing using mathematical models, he acknowledged that the sensitivity of the technology is also a limiting factor. Lab technicians, he noted, can also cause problems at any stage of the process. "Pooling involves more different work steps in the lab than individual testing," he said. "So it can be prone to human error."
One company that has sought to tackle the issue of sensitivity is France's Stilla Technologies. Earlier this month, the company, which is based outside of Paris, published a comparative study highlighting the use of its digital PCR technology for group testing of SARS-CoV-2 for pool sizes of up to 16 samples. The study involved nearly 450 samples and was run with partners at Bichat Hospital and Ecole Polytechnique in Paris using digital PCR and conventional real-time PCR. Scientists used Stilla's three-color digital PCR system, the Prism3, along with its Naica instrument, but the company recently introduced a six-color system, called Prism, which can also be used for virology testing.
"From day one, there has been an issue with pool testing and sensitivity," said CEO Remi Dangla. "When you have a person who doesn't have a lot of virus in their sample, and you dilute that low-concentrated sample, you dilute your virus even more and your probability of detecting just goes down," he said. "We knew we had a technology that had increased sensitivity over RT-PCR. So we thought, 'let's combine pool testing with our-high sensitivity.'"
The company and its partners published a preprint of the work in ArXiv earlier this month. "This could be the start of pushing this forward and making this a real tool for future diagnostics," said Dangla, adding that pool testing has been held back by the medical field because of the sensitivity issues. "We think, with our technology, it's a real weapon in the fight with this pandemic."
In his comments, Dangla noted that academic groups and companies alike have been busy developing new methods for pool testing for SARS-CoV-2. In recent months, for instance, researchers at Stanford Health Care Clinical Virology Laboratory published a paper in JAMA, and groups at the German Red Cross Blood Donor Service and Technion, the Israeli Institute of Technology, made announcements about new rapid pool tests. In May, a team from Technion described its approach in the journal Clinical Infectious Diseases.
Roy Kishony, a professor of life sciences at Technion and a corresponding author on the paper, said his team worked with partners at Rambam Health Care Campus on the project, which determined that a single positive sample could be detected in pools of up to 32 samples using standard techniques. The authors also reported that positive samples could be detected in pools of up to 64 samples, though this might require additional amplification cycles. They concluded that a pool test for COVID-19 would "allow expanding current screening capacities ... enabling the expansion of detection."
"This work opened the door to increasing the efficiency of COVID-19 testing by pooling together many samples in one tube," noted Kishony. "For a positive rate of 1 percent in the population, the approach can save over 90 percent of the PCR reactions," he said. "The approach can also be combined with combinatorial sampling where each sample is tested in multiple different pools."
Kishony said the group is now working on practical implementation of the approach to screen populations over time, as well as to monitor specific groups, such as army units, hospital staff, or factory workers. In terms of challenges, beyond gaining regulatory approval, carrying out pool testing requires systematic protocols and laboratory automation, he said. "It is much more of an execution challenge rather than a scientific challenge."
Regardless of the challenges, Max Planck's Eberhardt said that, once adopted, pool testing could "drastically improve the efficiency of testing" for SARS-CoV-2, especially when the prevalence of infection is low. He noted that, according to Chinese state media, labs were able to test 6.5 million people in nine days using a pool testing approach.
"Pool testing can be employed to test huge populations much more efficiently," noted Eberhardt. "Assuming a prevalence rate of 0.5 percent, our multi-stage scheme could test 10 times more people than individual testing."