The challenges to deploy diagnostic and surveillance testing for the COVID-19 pandemic will persist at least until a vaccine is ready if not well beyond that milestone. The state of COVID-19 testing as a regulated sector is a complex intersection of new and old technologies, questionable accuracy, availability hurdles, supply chain interruptions and problems with interpretation of results.

There are two types of diagnostics: molecular, which detects the presence of genetic material from the virus) and antibody (includes serology tests), which can identify people who previously had COVID-19. (See “The tests and supplies for COVID-19, explained,” for more details.)

In an effort to speed COVID-19 testing innovation and get the tests to market, the NIH’s $1.5 billion national COVID testing challenge invites entrants to compete for $500 million in direct funding.

Conversely, the FDA announced May 4 a reversal of its policy of leniency for serological tests, a change the agency said was necessary in part due to illicit claims made by developers. The policy went into effect immediately, signaling a shift in the agency’s approach to surveillance testing. Makers of serological tests now must forward validation data within 10 days of the test’s deployment, a sharp reversal of the prior policy, which allowed the tests to market without review to develop a picture of COVID-19 prevalence and incidence.

One of the issues regarding serological testing is reliability, which was taken up by a recent evaluation of a number of tests by the Covid Testing Project. This study evaluated 10 lateral flow assays (LFAs) and two enzyme-linked immunosorbent assays (ELISAs), and the researchers concluded that performance was heterogenous and that reader training is key to reliable LFA performance. Perhaps more troubling, however, was the conclusion that optimal use of serology will require evaluations of these tests across “the full spectrum” of disease stage, from asymptomatic cases to recovering patients, and that well-designed studies will prove vital to ensuring that serological testing can be relied upon for public health policymaking.

Molecular tests are still the only one-shot diagnostic

Molecular testing for diagnostic purposes initially drew the most interest in the early days of the pandemic and is still the only widely available test type that enjoys credibility as a one-shot diagnostic. Nonetheless, even molecular testing for viral RNA has its limitations as a diagnostic instrument.

Carmen Wiley, president of the American Association for Clinical Chemistry, told BioWorld that polymerase chain reaction (PCR) testing for viral RNA can return a false positive for active disease if the patient has recovered, but still bears fragmentary RNA in the sampled tissue. She said false negatives are still seen as primarily driven by pre-analytical issues, such as poor sample quality or degradation in shipping. The term of sample degradation varies, however, depending on whether the sample is shipped in refrigerated containers, but refrigeration is not the only difference between specimen kit requirements.

The question of antibody testing via serology seems to hinge on the variable production of the three main immunoglobulins of interest, immunoglobulin A (IgA), immunoglobulin G (IgG) and immunoglobulin M (IgM). Wiley confirmed that the conventional thinking about the appearance of immunoglobulins is that IgM and IgA are detected earlier than IgG, which is more persistent. However, she said recent data seem to suggest that IgG emerges almost as quickly as IgM, although this finding has not been exhaustively corroborated, and consequently, “we’re still evolving our understanding of this particular virus.”

Despite the inclination to think that a test directed solely to IgG might make sense if these findings prove out, Wiley said most assays on the market are total antibody serology tests and thus single-antibody testing is not a practical approach. “Until we have more experience, we won’t know exactly how valuable it is to break them out individually” in a multi-antibody assay either, she said.

Wiley, who is also the chief medical officer at Veravas Inc., of Charleston, S.C., said it is difficult to pin down exact figures for the number of labs that possess a particular type of testing equipment, but noted that antibody assays can be run on low-volume benchtop units as well as high-volume systems that can process several hundred samples per hour. These systems might rely on clinical chemistry and/or immunochemistry testing techniques, and Wiley said that while the installed base of instrumentation determines the type of test a lab can run, smaller labs may rely fairly heavily on manual methods whereas reference labs are more likely to possess high-throughput analyzers.

When asked what she sees as the most pressing concerns about antibody testing, Wiley said the interpretation of positive test results is still at the top of the list in her view. “The timing of collection and prevalence of the disease in the population really affects the interpretation,” she said. If a serology test is widely deployed in a small town with a prevalence of only 5%, a test with a 95% sensitivity and specificity might sound more than adequate, but she explained, “when you apply that serology test to this population, the positive predictive value is 50%.”

Wiley noted that this scenario also suggests a negative predictive value of 99%, so a negative result may be reliable regarding exposure. She said it’s critical that this be understood by the public, adding that when the prevalence is quite high – as in New York – a serology test is much more likely to render a true positive result. Consequently, there is a need to interpret serology test results with caution as the economy gradually reopens.

Wiley said a clearly symptomatic patient who is negative for a serology test may be referred for molecular testing to determine that patient’s status. The likelihood of obtaining a molecular test is somewhat dependent on local conditions, however, such as test availability and the screening approach adopted by state and/or local authorities.

One of the key issues is availability of supplies, but the goal over the next three to four weeks is to have sufficient testing capacity – including supplies, sample and test kits and reagents – to accommodate all people who are referred for testing. A physician’s order is still necessary for a test most of the time.

There is a fair amount of discussion of next-generation sequencing (NGS) as an alternative to PCR for diagnostic molecular testing, but Wiley noted that relatively few labs have the capacity to run NGS testing. An NGS system also offers more throughput than is needed to evaluate a viral RNA sample, but Wiley suggested that stakeholders stay tuned as a debate may emerge in the coming weeks as to the role, if any, for NGS diagnostic testing.

One of the less-discussed dilemmas is staffing at clinical labs. Wiley said staffing is still sufficient because of the decline in elective procedures but when the volume of elective procedures resumes a historically typical pace, “those lab personnel are going to have to go back to their regular areas.”

Lab personnel certification is largely regulated by the states, and Wiley said the number of lab personnel who are qualified to handle communicable disease testing may be insufficient. Whether a lab professional can be quickly cross trained to communicable disease testing depends on that state’s regulatory regime. Some states require a formal training program and certification, “so getting someone ramped up quickly … is nearly impossible.” Other states allow a lab employee with a bachelor’s degree in that scientific area to undergo a less rigorous training and documentation regime, allowing that individual to fill in for the time being.

Data vital, but volume was critical in early going

Peter Pitts, president and co-founder of the Center for Medicine in the Public Interest, told BioWorld, that while a number of serological tests were on the market without any review from the FDA, “for us to move forward through COVID, data must be our ally. The only thing worse than no data is bad data,” he said.

Pitts, who previously served as the FDA’s associate commissioner for external relations, said he is concerned that there will be more false positive results than false negatives from serological testing, which creates difficulty for policymakers charged with making decisions regarding public health and safety. That said, Pitts said he might have steered the same course the FDA did because “volume counts” when it comes to developing an understanding of the regional, national and international impact of the pandemic, and dilatory regulatory requirements cannot be allowed to stand in the way.

Still, it is vital that any decisions be made on hard data, and Pitts said, “that’s a lot harder than it sounds” in this predicament. Regarding the respective importance of PCR vs. serology tests, he noted that some of the thinking behind serology testing in the early going was that the antibodies could be used to aid in the effort to develop convalescent plasma treatments. Another complicating factor for serology testing is the question of whether a recovered patient can be re-infected, which could reduce the utility of serology testing.

PCR tests identify infection rate

PCR testing is critical in part because it paints a more accurate picture of the infection rate, including the rate of asymptomatic infections to overall infections, Pitts said, adding that such information is going to be critical as the economy inches into a greater or lesser state of reopening. He said a vaccine could arrive within 10 months to a year, but he noted that the flu vaccine is short of utterly reliable, which will confound any efforts to discern which test a patient should receive. Physicians and other health care professionals may have been inclined over the past few months to advise any symptomatic patients to stay home on the prospect that the symptoms were COVID-19 and not the flu, but that might have been good advice for the flu as well, given the risk of co-infection.

The U.S. is working to move from mitigation to containment, and Pitts said policymakers must strike a difficult balance between test quality and quantity as the move towards containment gains momentum. The FDA, CDC and the NIH’s National Institute for Allergy and Infectious Diseases are collaborating to validate serological tests, but Pitts said it is not clear that NIAID can validate the tests with sufficient speed. Pitts said the NIAID is not being particularly transparent about the validation, either, adding that it is not clear why the FDA is not allowing public and private labs to handle these validation review tasks.

Beyond the FDA

Regulatory agencies across the globe have been forced to adjust their expectations of test makers in the COVID-19 pandemic, and while the U.S. FDA has drawn many headlines, it is by no means the only agency that has broken with regulatory tradition in 2020.

The U.K. Medicines and Health Care Products Regulatory Agency posted a series of guidances and policy documents related to diagnostics and test kits, including efforts to scale up testing programs. As is the case with the FDA, MHRA said there are no home tests that enjoy regulatory approval. MHRA also announced May 6 that a new reagent is available for use in the U.K. to address the running shortage of reagents.

Australia has also acted to expand testing for the pandemic along with an effort to roll out a national surveillance app, while Australia’s Therapeutic Goods Administration had already provided an expedited regulatory review process for a variety of devices. Health Canada has responded with its own expedited process for testing and other COVID-related needs.

Editor's note: This article is part of a series of stories assessing the state of the rapidly evolving COVID-19 diagnostics.