Pushing limits of detection for the new coronavirus
BioCentury is providing this story for free given the urgent need for information about the COVID-19 crisis. For more analysis, sign up for our daily email.
Even as companies strive to create ever-more sensitive diagnostics for COVID-19, it’s still not clear what limits of detection are required to determine whether people are truly in the clear.
The discovery that asymptomatic COVID-19 patients can transmit the virus means it is essential to find out who carries the virus in the absence of signs of disease (see "Reopening the Nation will Depend on Diagnostics").
COVID-19 tests fall into two broad buckets: assays that diagnose active infections by detecting viral nucleic acids in nasopharyngeal swabs; and serological assays that indicate whether someone has been exposed to the virus by detecting antiviral antibodies in blood samples. Only the former can be used to diagnose active COVID-19 and guide clinical care, though the latter are being positioned as a way to determine whether someone has protective immunity, and can therefore go back to work.
To know who is capable of spreading the virus and intervene before severe disease sets in, diagnostics that measure viral load need to be sensitive enough to allow infections to be caught early, as well as to dictate whether people who no longer show COVID-19 symptoms should be released from the hospital.
The problem is that there’s no consensus about what level of sensitivity is the goal.
Companies have jumped in to develop tests based on conventional platforms and newer technologies that are faster, or more informative (see "COVID-19 Diagnostic Tech Tableau" and "COVID-19 Spurs Rapid Diagnostics").
Viral nucleic acid tests granted Emergency Use Authorization (EUA) by FDA to detect SARS-CoV-2 exhibit a wide range of limits of detection (LoD), defined as the lowest amount of the target that the test can detect, at least 95% of the time. A lower LoD indicates a more sensitive test (see Table: "LoDs for FDA-Authorized COVID-19 Diagnostics").
"What's the definition of 'cured'? Detection sensitivity is critical."
Jinzi Wu, chairman and CEO of Ascletis Pharma Inc. (HKEX:1672), said experience with HCV and HIV suggests defining a LoD that acceptably minimizes false negatives requires large, extensive clinical studies that could take years to complete.
"What's the definition of 'cured'? Detection sensitivity is critical," said Wu. "When you say somebody is ‘cured’, the definition in laymen's terms is: in this person, you could not see the virus, and that person is off drugs."
According to Wu, insufficiently sensitive tests were responsible for the false perception that interferon therapies cured HCV patients in the 1990s; the infections flared up again after treatment withdrawal. Improvements in tests detecting HCV load, and studies that monitored patients for several years, were key to supporting the approval of HCV-curing therapies, he said.
Wu added that long clinical studies of HIV patients were key to identifying the viral titers below which there is very low transmission to others, and the corresponding sensitivity of tests needed to evaluate someone's transmission status.
"It's pretty hard for us to define right now, if you have a certain number of copies of coronavirus, you are in a high transmission group," Wu said. "Eventually we will know, just like we did with other major viruses."
The acute nature of COVID-19, and the massive mobilization of research the crisis has provoked, could lead to a speedier understanding of the relationship between viral load, recovery and and transmission than in previous epidemics. In the meantime, some new entrants to the U.S. market are pushing sensitivity limits lower.
Becton Dickinson and Co. (NYSE:BDX), whose COVID-19 test received an EUA on April 2, was the first among the FDA-authorized tests to cross the 100 viral genomic copies/mL LoD barrier, at 40 copies/mL. And Bio-Rad Laboratories Inc. (NYSE:BIO) is seeking an EUA for its droplet digital PCR (ddPCR) test, which exhibited an LoD of 21 copies/mL in a Wuhan University study.
The ddPCR platform owes its greater sensitivity to several technical factors including the lack of reliance on standard curves, which are a major source of error for quantitative PCR tests, said Bio-Rad life sciences spokesperson Carolyn Reifsnyder.
Table: COVID-19 diagnostic tests with EUA
More than 25 diagnostic tests have received Emergency Use Authorization (EUA) by FDA for COVID-19 testing in the last two months. The limits of detection (LoD) for the virus range from 40 copies/mL to 100,000 copies/mL; five tests use the measure copies per reaction, and range from three copies to 200 copies per reaction.
Two use TCID50 (median tissue culture infectious dose) per mL, and have very close values at 0.009 and 0.01 TCID50/mL. TCID50 represents the viral load at which 50% of cells are infected when a solution containing the virus is added to cell culture.
An assay's LoD is the lowest amount of the target that the test can detect at least 95% of the time. A lower LoD indicates a more sensitive test.
LoDs reported as viral genomic sequence copies per volume are primarily reported in copies per mL; asterisks denote LoDs originally reported as copies per μL.
The CDC reported separate LoDs for two versions of its test involving different PCR primers.
Most of the diagnostics that have received EUA from FDA are reverse transcription polymerase chain reaction (RT-PCR) tests; the ID NOW COVID-19 test from Abbott Laboratories (NYSE:ABT) is based on non-thermocycling probe technology (see "COVID-19 Diagnostic Tech Tableau").
FDA has also granted EUA to a serological test from Cellex Inc. that detects patient antibodies against COVID-19; because these tests target variable human proteins, they are evaluated by their clinical sensitivity, rather than an analytical LoD. The Cellex test has a sensitivity of 93.8% and a specificity of 95.6%.
Source: Manufacturers' instructions for use documents