The little over six months since Dr Tedros, the Director General of the World Health Organisation, famously urged countries to “test, test, test” have seen an explosion in the number of assays available for detecting the COVID-19 virus. At the end of March, there were only 24 assays authorised by the FDA for emergency use; this figure has ballooned to over 190, of which the vast majority—about 160—are PCR tests. Only seven have been approved for rapid antigen.
So far, PCR tests are the only tests that have been universally embraced by national health systems, and several territories now require the presentation of a negative PCR test result before entry is permitted—with some using this to determine whether an individual needs to undergo quarantine. PCR tests are, in short, the gold standard for the diagnosis of an acute infection.
This article explains why this is, introduces other types of test and the notion that each one has different strengths and weaknesses, and raises the matter of whether this gold standard is really required all the time.
A primer on types of Covid-19 tests
Tests for COVID-19 can be divided into two categories: tests that detect antibodies, and tests that detect the virus.
Antibody tests, also known as serology tests, look in blood samples for the presence of antibodies produced by the body specifically to fight off the virus. While these tests have some value in the diagnosis of ongoing infection, especially in areas where PCRs are simply not accessible, too costly or results taking too long, they are most frequently used for identifying past infection, such as in seroprevalence surveys.
Tests that detect the virus are used to look for current infections. Of these, the ones most widely used are nucleic acid amplification tests (NAATs)—also known as molecular tests—and antigen tests. While both types of tests look for the virus in samples taken from the respiratory tract, they do so in very different ways.
PCR is an example of a NAAT, where the sample is prepared in a way that isolates the genetic material (RNA) of any virus present; the genetic material is then amplified to allow the test to detect certain genetic sequences. NAATs can be further categorised by whether they can be performed at the point of care (point-of-care tests, or POCTs) or if they require a laboratory with complex equipment; PCR tests are of the latter type.
Antigen tests, on the other hand, contain labelled antibodies (proteins) that bind to any SARS-CoV-2 virus present in the sample to form conjugates; it is these conjugates that are detected by antigen tests. Antigen tests currently available are generally rapid tests that take the form of dip sticks or cartridge casettes, with some requiring the use of small, toaster-sized analysers to interpret results. General characteristics of these tests are summarised in the table below:
PCR tests and the gold standard
The main reasons why PCR is often described as the “gold standard” of testing are because this type of test is:
- Highly sensitive (few false negatives) and highly specific (few false positives); we know this both due to the principles and mechanisms by which such tests operate, and from laboratory and clinical evaluations. However, this is highly dependant on the operator’s skills.
- Very low limits of detection, meaning that they can detect the presence of even a minuscule amount of viral RNA. They can thus rightly be described as the gold standard method for the detection of viral genes in a sample.
PCR tests also fill the role as a reference standard. In order for a test to be evaluated, it should ideally be compared with a test known to be 100% accurate; since no such test exists, the next best thing is a reference test that meets certain criteria in reliability and accuracy — a role filled by PCR tests.
In clinical evaluations of POCTs and rapid antigen tests, the accuracy of a test is determined by how frequently its results agree with those from a PCR test.
Different Objectives, Different Tests
When Dr. Tedros urged countries to test more frequently, the objective he had in mind was to “[break] the chains of transmission”; testing should be used to provide actionable information, on whether someone should self-isolate or to inform contact tracing.
But is PCR testing really the best tool for this?
One reason that SARS-CoV-2 is so much more infectious than SARS-CoV-1, the virus causing SARS, is the way viral load—how much virus there is—in the upper respiratory tract changes over the course of the illness. In SARS-CoV-1, viral load is highest during the second week of illness, whereas in SARS-CoV-2, viral load is highest in the days preceding and week following symptom onset.
As there is considerable evidence that there is a strong correlation between upper respiratory tract viral load and infectiousness, individuals infected with SARS-CoV-1 showed symptoms about a week before they are most infectious, whereas those infected with SARS-CoV-2 are already highly infectious before symptom onset.
Testing only symptomatic individuals with COVID-19 thus has limited effect on breaking the chains of transmission; by the time symptomatic individuals get tested, not only have they likely been highly infectious for days already — making contact tracing difficult — but it is also likely that they will not be infectious for much longer (usually 8-9 days after symptom onset).
This means they will self-isolate for far longer than is necessary (currently 14 days). Harvard epidemiologist Michael Mina, M.D., PhD., estimates that fewer than 5% of symptomatic people are tested in time to prevent them from transmitting the virus to others.
The only way to ensure that asymptomatic and pre-symptomatic cases are identified in time for contact tracing to be meaningful is thus to perform screening tests as frequently as possible, testing the entire population of a school/workplace/country several times a month, if not several times a week.
While this sort of testing is not impossible with PCR, the costs and logistical challenges make it prohibitive, at least with the current technology.
Sensitivity and Infectiousness
The very high sensitivity of PCR tests to the presence of viral RNA could arguably be a problem in managing the pandemic as it may be ‘too sensitive’.
A patient can test positive in a PCR test long after she is no longer infectious: the average number of days after symptom onset that RNA can be detected is 17 (in one case, the maximum known number is 83), but it’s rare for the virus to be viable and transmissible for more than 9 days after symptom onset.
If population screening is performed infrequently or as a one-off event, cases that PCR tests identify are far likelier to be individuals who are no longer infectious than those who are asymptomatic or pre-symptomatic.
Not only does this mean that many non-infectious or recovered people are forced into quarantine, but also that resources are wasted in tracing contacts often after their period of infectiousness has passed. As Mina puts it, positive PCR results can be false positives for actionable results.
POCTs and Rapid Antigen Tests for Screening
The cost per test of POCTs and Rapid Antigen Tests is a mere fraction of that of PCR tests, and they have considerably lower resources and skill requirements. This enables them to be deployed at scale in a wide range of environments, allowing for far more frequent and widespread screening of asymptomatic individuals. This maximises the probability of identifying early stages of COVID-19 on-site, without the complex logistics of PCR.
In October, the supply of rapid antigen tests skyrocketed to more than 150 million, six times the number of tests of any kind performed in August in the United States, according to the COVID Tracking Project. This is because the technology can be scaled up quickly and cheaply to meet demand, unlike with PCR.
POCTs and Rapid Antigen Tests are often cited for not being sensitive enough—but, as explained above, the very high sensitivity of PCR tests can also be a problem.
Very high sensitivity is, moreover, not necessary for identifying the most infectious individuals, which is the objective of frequent screening. In these individuals, the virus is present in such high quantities that even a less sensitive viral test should be able to detect it.
It should also be noted that during the early stages of infection, the virus multiplies at such a rate that the time difference between a positive PCR result and a positive rapid antigen result is very small – as little as four hours, and definitely less than a day. This difference is effectively offset by the much higher frequency of rapid antigen testing. In short, test sensivity is secondary to frequency and turnaround time.
The frequent use of a rapid low-cost POC test has an important role to play in disrupting transmission chains of individuals who are potentially infectious. Professor Mina estimates that the U.S. alone will need tens of millions of tests daily if schools and businesses are going to stay open without causing a surge in cases.
An additional benefit Rapid Antigen Tests have is that an individual who is no longer infectious is highly unlikely to test positive, as genetic fragments of lysed virus do not react with the antibodies in such tests.
PCR tests are not ideally suited for asymptomatic screening, as they are too expensive to perform in the frequency needed to catch cases when they are at their most infectious. Moreover, their very high sensitivity, together with their inability to determine whether the virus in a sample is viable, means people who are not at risk of infecting others are forced into self-isolation.
POCTs and antigen tests, on the other hand, can be systematically deployed at scale and at a much greater frequency, which more than makes up for their slightly lower sensitivity.
Rapid antigen tests should be deployed in conjunction with PCR tests in the fight against the pandemic. Just last week, the European Commission urged member countries to use rapid antigen tests for the diagnosis of COVID-19, to scale up testing capacity, ensure a common approach and more efficient testing strategies: “Member States are encouraged to conduct rapid antigen tests in addition to RT-PCR tests to contain the spread of the virus, detect infections and limit isolation and quarantine measures.”
In conclusion, PCR tests are indeed gold standard tests, especially when it comes to diagnosing an individual who is symptomatic. However, for the purpose of snuffing out an outbreak, POC NAATs and more pointedly Rapid Antigen Tests are the way to go.
 Cevik, M., Kuppalli, K., Kindrachuk, J., & Peiris, M. (2020). Virology, transmission, and pathogenesis of SARS-CoV-2. BMJ, m3862. https://doi.org/10.1136/bmj.m3862
 Basile, K. et al., (2020). Cell-based culture of SARS-CoV-2 informs infectivity and safe de-isolation assessments during COVID-19. Clinical Infectious Diseases. https://doi.org/10.1093/cid/ciaa1579