With molecular testing capacity to identify people with active virus ramping up quickly (increasing 100-fold since March 10th) and blood tests anticipated for antibody testing on the horizon, multiple government officials are now indicating that testing may allow some to return to work in May or June. Moreover, broader population testing may allow more effective and selective quarantining. There’s increasing reason to believe that "silent" COVID-19 cases—those who are infected but have not experienced symptoms, or those who have recovered from mild symptoms and were never tested—are significant.

 

While a vaccine holds the promise of protecting those who have never been exposed, it’s not likely to be widely available until late 2021 or 2022. However, antibody testing may allow people to return to work starting in Q3 and Q4 of 2020, in addition to protecting essential workers and allowing them to stay on the job safely. While there’s promise in antibody testing, we believe that both testing for active virus and antibodies is essential, and that companies and laboratories need to be purposeful in how these tests are brought to market to maximize impact.

 

 

Testing for active virus has relied primarily on molecular tests that amplify strands of virus RNA that are specific markers of the SARS-COV-2 virus that causes the COVID-19 infection. Recent research from SVBLeerink estimated that testing volumes have expanded from approximately 1,000 per day on March 10th to 100,000 tests per day on March 26th, including rapid-point-of-care molecular tests for detecting active virus from companies like Abbott, Becton-Dickinson and others. Patients test positive when they have active virus and will test negative for active virus when they have recovered. Consistent with WHO guidelines, some countries have considered two consecutive negative tests given 24 hours apart as the standard for defining a “recovered” patient.  

 

The body’s immune system produces two antibodies when fighting the infection: immunoglobulin-M or IgM after several days, and then later immunoglobulin-G or IgG. IgG may confer protection and immunity for recovered patients and thus prevent them from being reinfected, though the level of antibodies necessary for protection and the duration of protection are not yet known. Thus, testing for antibodies may allow for the identification of patients who have already been exposed and recovered from the virus whether or not they experienced symptoms or received a diagnosis.

 

The first FDA Emergency Use Authorization (EUA) for an antibody test was granted to Cellex, and the assay can detect IgM and IgG antibodies to the coronavirus. Of the 32 EUAs for SARS-COV-2 diagnostic tests listed on the U.S. FDA website as of April 8th, Cellex is the only antibody test, and the remainder are molecular tests for active virus. The Foundation for Innovative New Diagnostics (FIND) is tracking tests currently available and in development for COVID-19, and their list includes 233 antibody tests available or in development globally as of the date of this article. Unlike molecular tests, however, antibody tests may be marketed in the U.S. without an FDA Emergency Use Authorization once manufacturers have performed appropriate evaluation to determine if the tests are accurate and reliable according to FDA and state-issued guidelines. Although antibody testing for COVID-19 is still in the early stages, there are already many antibody tests in use today in the U.S., and many more in the validation stage that will be on the market in coming weeks and months. These will include tests focusing on high-throughput volumes for central labs as well as point-of-care options.

 

 

Data from multiple sources has suggested that the number of silent and undiagnosed cases is likely significant—particularly so in the U.S., where tests were initially reserved for those who were the most symptomatic. The U.S. CDC indicated that as many as 25% of people infected with the virus may not show symptoms. A Hong Kong team suggests that 20 to 40% of transmissions occurred in China before symptoms appeared, and as many as 18% of people infected on cruise ships never developed symptoms. A population study in Iceland across age groups suggests that rates of infection in younger patients (who may be asymptomatic spreaders) is significantly higher than realized, and as many as 50% of infections may be asymptomatic. Many people in the hardest hit areas who had mild symptoms and were not tested but have since recovered will also add to the uncounted but “recovered” population. And the number of diagnosed and documented recovered patients also grows with every day that passes.

 

Even after the number of new cases and death rates peak in particular cities and states, there will be a significant risk of follow-on waves of outbreaks if social distancing is not maintained for those who have not developed immunity. This threat of follow-on waves coupled with the need to quickly return to normalcy requires a different and more robust testing approach.

 

 

Validation of antibody tests requires testing and documentation with respect to specificity (avoiding false positives and correctly identifying patients who are negative as negative) and sensitivity (avoiding false negatives). Sensitivity is more complicated typically as antibodies build over time after a body initiates an immune response. Sensitivity of antibody tests in COVID-19 patients has been reported to be much lower in the early days after an infection than in later stages of an infection. Thus, a false negative rate may be higher for some patients than for others even for a particular test, depending on when that test is administered. For this reason, we anticipate that administering both a molecular test—which ensures there’s no active virus that could be in the early stages of an infection—and an antibody test, which may confer immunity, may be necessary. 

 

Furthermore, many antibody tests may not indicate the level of antibodies present—just their presence—and it’s not yet known whether there is a minimum level of antibodies necessary to confer protection and immunity from reinfection for a minimally acceptable period of time. People with antibodies for severe acute respiratory syndrome (SARS) had immunity that typically persisted for two years and declined by three years. However, there’s also a possibility that the SARS-COV-2 virus could mutate, perhaps seasonally as is the case with the influenza virus.  

 

There have been sporadic reports from China and South Korea of the potential for reinfection, or possibly reactivation of the virus, from patients who had recovered from COVID-19 and tested as negative for active virus. However, these reports could be confounded by testing accuracy in the original diagnosis or subsequent test, and represent a tiny fraction of people compared to the population of recovered patients that grows daily.

 

 

In order to return to work, patients who have not received a diagnosis previously may need two tests: a test for active virus to ensure they are not currently infected, and a test for antibodies that would signal potential immunity. Those with no active virus and with the SARS-COV2-2 IgG antibodies may then be given permission to go back to work (perhaps with a certificate and being monitored as part of a registry to track durability of protection and any reinfections from either current or future mutated strains of the virus). Patients that were tested and diagnosed with COVID-19 may only require an antibody test (and eventually there may also be presumptive immunity conferred for those who were diagnosed and have been symptom-free and recovered for a sufficiently long period of time). Beyond broad availability of testing, there are a few key considerations for how these tests are designed, deployed, accessed and produced.

  • Design considerations: Oftentimes, developers of tests face a trade-off between specificity and sensitivity, meaning that if you make a test that’s great at correctly identifying people who have the antibody then it may become worse at correctly identifying people who don’t have the antibody. The opposite is also true. While overall accuracy is important (high sensitivity and specificity), different types of tests can calibrate on these dimensions based on the needs for that particular test. In this situation, the need for antibody testing is to safely identify a population that can return to work. Thus, if the test indicates that antibodies are present, you want to be sure that such an indication is correct. It would be better to keep people sheltered than expose someone unnecessarily to the virus. Therefore, in this situation a higher calibration for specificity would be preferred and a minimum level of both specificity and sensitivity should be sought to ensure consistency across laboratories (for example, greater than 95%). 
  • Deployment considerations: The level of testing required for an entire population is beyond existing capacity levels and thus deployment needs to be judicious and targeted. Geographic areas that have been hit especially hard should be targeted first, and jobs that are critical to everyday function and are at higher risk of exposure (such as healthcare, police, supply chain, etc.) should be supported through broad-based testing. While centralized laboratories would help to ensure consistent performance standards, decentralized (or testing conducted by the provider at site of care) has the benefit of faster turnaround time. In this particular situation, the standardization of high performance standards outweighs the benefit of one to four days faster turnaround time (whereas for initial identification of COVID-19 patients based on active virus, turnaround time may be much more important).
  • Access and pricing considerations: Immunoassay testing for antibodies is likely to be significantly less expensive than the PCR testing utilized for active virus (though pricing will also vary by type of testing technology and other factors). Typically, testing is priced through codes knowns as Current Procedural Terminology, or “CPT” codes, and reimbursed on a per-unit basis. Addressing this challenge calls for contracting with government, provider/payer systems, and employers to broadly support the targeted population described above. As an example of employer-based initiatives, Amazon has announced that they will begin building incremental testing capacity in order to test warehouse employees. Many other large employers should also be eager to follow similar measures to ensure the safety of essential workers still on the job, and eventually to allow those with immunity to return to work. 
  • Production considerations: Early on, our virus testing faced challenges due to lack of raw materials needed to conduct such tests. Given the requirements for broader-based antibody testing, a similar situation may arise. We need a coordinated effort on the supply chain such that demand and supply is better calibrated, unproductive inventory levels are minimized, and supply bottlenecks are reduced. Right now, most testing needs are determined locally and thus sub-optimized when considering the broader population.  

Returning to work will likely be a drawn-out and potentially individually and geographically varied process. It’s not yet known how antibody testing would be rolled out or paid for, though a combination of government and private companies participating is likely. Rationing of scarce capacity and ethical issues related to treating those with immunity differently than others would also need to be resolved. Along with clinical trials of existing and new drugs, therapies and vaccines that may help treat the sick and curb the spread of COVID-19, antibody testing will be an important part of the solution that is likely to be increasingly available this year.