The evolving zeitgeist of COVID-19 testing
What paradigms could stick, and what questions remain, for COVID-19 testing one year in
What paradigms could stick, and what questions remain, for COVID-19 testing one year in.
Becoming an essential pillar of the pandemic response has pushed the diagnostics industry toward more rapid and decentralized technologies, with a greater focus on routine asymptomatic screening.
One year into the COVID-19 crisis, an array of testing technologies have been developed, scaled up and deployed around the world. But questions remain about what test performance is good enough and how results should be used, particularly for individuals outside of institutional testing programs.
The first half of 2020 saw a scramble to ramp up capacity for the gold standard method of detecting RNA viruses in symptomatic patients — lab-based reverse transcription polymerase chain reaction (RT-PCR) — coupled with a focus on serological immunoassays for antibodies against SARS-CoV-2, to determine who had already built up immunity against the virus and could safely meet with others.
But by the second half of the year, it became clear that the utility of serological tests is limited by lack of understanding about what type and level of antibody response constitutes protection. The center of gravity shifted toward routine asymptomatic screening as the key to in-person gatherings, with a growing role for rapid, point-of-care technologies like antigen tests.
While rapid point-of-care testing modalities had been developed for other viral infections including influenza and Zika, interest generally waned as outbreaks wound down. The technologies had marginal standing compared with lab-based RT-PCR, which has a decades-long track record of high sensitivity and specificity, and hardware widely available in labs across the world.
But the scale and duration of the COVID-19 crisis, and the flood of R&D, government investment and product authorizations it has produced, now stands to make rapid, point-of-care or at-home testing a central and permanent fixture of the diagnostic landscape, particularly if issues regarding incentive structures, reimbursement and patentability of tests are addressed.
The prevalence of asymptomatic SARS-CoV-2 infections and their major role in viral spread has also forced the healthcare ecosystem go beyond its historical focus on diagnosing symptomatic patients and open the door wider to mass screening of seemingly healthy individuals.
The shift coincides with the maturation of liquid biopsy technologies for early cancer detection, which offer a template for balancing the risks and benefits of mass screening, and demonstrating its value.
Yet the embrace of asymptomatic screening and rapid point-of-care tests, which are generally less sensitive than lab-based assays, has raised questions about the accuracy of results, and how individuals interpret, report and act on them, particularly as COVID-19 testing moves into the at-home and over-the-counter setting.
A major open question is the viral load threshold for transmission, which would give developers a clear benchmark for what test sensitivity is required to flag infectious individuals. Such thresholds historically take years to establish.
Tests authorized to detect SARS-CoV-2 have a wide range of sensitivities, and the dramatically shortened development timelines necessitated by the crisis have led to some stark gaps between manufacturers’ initial characterizations and test performance in the real world.
Challenges include the fact that viral load, and thus test sensitivity, varies throughout the course of infection, parameters that are especially hard to pin down in the absence of symptoms.
The biggest concern is the risk of false negatives from insufficiently sensitive tests, which could lead infectious individuals to congregate with others and possibly relax preventive measures like mask wearing and social distancing.
While individual lab-based RT-PCR tests, particularly those with higher limits of detection (LoD), come with some risk of false negatives, the risk is greater for pooled RT-PCR tests and rapid, point-of care tests, whose lower cost makes them more amenable to mass screening.
Those risks are mitigated in populations tested frequently as part of institutional screening programs, which can quickly catch cases that were previously missed, but remain acute for individuals using an isolated result from a single point in time as a basis for decision making.
Overcoming the risk of harm from false negatives will require expanding access to frequent testing regimens, and creating clear consensus guidelines about how to interpret and act on results — for example, the window of time a result is likely to be valid for and protocols for isolation and re-testing.
In the absence of a known viral load threshold for transmission, an ongoing debate is whether less sensitive rapid tests are better at pinpointing who has a high enough viral load to actually be infectious.
A CDC study published Tuesday reported that the BinaxNOW rapid antigen from test Abbott Laboratories (NYSE:ABT) had sensitivities of 62.4% and 35.8% in symptomatic and asymptomatic individuals, respectively, compared with an RT-PCR test. In the subset of symptomatic and asymptomatic individuals whose samples were capable of infecting cells in culture, the respective sensitivities were 92.6% and 78.6%.
Another question is whether delivering quantitative results for RT-PCR tests, instead of a yes or no answer, would help assess individuals’ likelihood of transmitting the virus or getting sick.
A quantitative result could come in the form of a CT value, the number of temperature cycles required to amplify viral RNA levels above a detectable threshold, which reflects the amount of virus present in a sample. The utility of CT values is complicated by variability in RT-PCR systems, and in how respiratory samples are collected by different individuals.
Without the luxury of time to optimize test parameters, public health officials need to develop frameworks for continuously evaluating and maximizing the utility of tests as more data rolls in, ideally via a harmonized approach that gets separate agencies such as CDC, FDA and HHS on the same page.
Regulators have had to adapt to a rapidly shifting COVID-19 testing landscape in real time. The year has seen FDA recalibrate whether to review serological tests, whether frequent testing can compensate for lower sensitivity, and how much to encourage off-label asymptomatic use of tests authorized for symptomatic individuals.
But they have also sown confusion and concern by sending conflicting messages on how to tests should be used and regulated, and by scaling up use of technologies whose performance later proved lacking.
A recent example of the latter includes a test from Curative Inc. that was used to screen members of the U.S. Congress and conduct large public testing programs, including in Los Angeles County, a pandemic hotspot. On Jan. 4, FDA issued a letter warning that the test had a high risk of false negatives, particularly when used outside of its stated indication of symptomatic individuals within 14 days of COVID-19 symptom onset, as it was in the Congressional screening program.
Rapid tests have been a central focus for both the outgoing and incoming U.S. presidential administrations.
According to a fact sheet released Nov. 24, the Trump Administration gave $10.25 billion to states, tribes and territories for testing, $20 billion to nursing homes for testing and other needs, and $847 million to support testing of uninsured people. The document highlighted spending on rapid, point-of-care tests, including $760 million for Abbott’s BinaxNOW test.
On Jan. 14, the incoming Biden Administration announced a plan to invest $50 billion to expand testing by purchasing rapid tests, growing lab capacity and helping schools, local governments and institutions such as prisons and long-term care facilities implement regular testing protocols, with the goal that any American can get a test for free if they need one.
Expanded government investment in routine rapid testing, and growing public expectations that such tests should be available, could help break the dominance of lab-based testing in the U.S.
Next diagnostic needs
As COVID-19 vaccination picks up steam, tests that reflect whether individuals have developed neutralizing antibodies and T cell responses against the virus are gaining renewed importance as tools to identify correlates of protection, which will be key for understanding how long protection lasts and establishing surrogate endpoints for the next wave of vaccine candidates.
An important consideration is that the COVID-19 vaccines authorized so far induce immunity against the SARS-CoV-2 spike protein, while many serology tests detect antibodies against other antigens such as the virus’ nucleocapsid protein.
This mismatch could create anxiety in vaccinated individuals who undergo serology testing if the results are not properly contextualized by healthcare professionals, but it also provides an opportunity to assess the effect of vaccines on asymptomatic transmission.
Moderna Inc. (NASDAQ:MRNA), for example, has disclosed plans to study the rate of asymptomatic cases in its trial population using serological assays that detect antibodies against the virus’ nucleocapsid protein, which are more likely to have been induced by an infection than by the vaccine.
The emergence of more virulent strains of SARS-CoV-2, and the looming threat of the next pandemic, have also made clear the need for continuing investment in pathogen-agnostic surveillance technologies based on next-generation sequencing.
One emerging way of flagging new variants like B.1.1.7 that have mutations in the virus’ spike protein is by sequencing “S gene dropout” samples— those that test positive for other SARS-CoV-2 genes via RT-PCR, but not the gene that encodes the spike protein.
Illumina Inc. (NASDAQ:ILMN) and Helix Opco LLC, a population genomics company administering COVID-19 tests, are collaborating to track the emergence and prevalence of novel SARS-CoV-2 strains, with support from the CDC; the partners have used the S-dropout method to catch its first set of cases.