Blog 9: COVID-19 Testing

At this point, everyone knows that testing is available to determine if someone is currently infected with SARS-CoV-2, the coronavirus responsible for the COVID-19 pandemic. These tests involve taking a sample of “nasopharyngeal secretions” (i.e. usually sticking a swab in the nose to take a sample from the back of the throat), then adding reagents to the sample to prep and isolate the RNA present. Scientists then perform RT-PCR (reverse transcription polymerase chain reaction) to determine the presence of known viral RNA sequences. PCR is essentially a way to amplify the genetic material present in a given sample, which allows us to detect viral RNA in very small samples. Because coronaviruses are RNA viruses, there must be an additional step of reverse transcribing the RNA into cDNA (complementary DNA), due to the limitations of PCR. RNA is far less stable in the high temperature conditions required for PCR (and just in general), and the enzyme used (Taq polymerase) is a DNA polymerase that can’t replicate RNA. Hence, the addition of reverse transcriptase. Now, all of this science sounds good and dandy, but there are some drawbacks to this genetic testing.

Genetic testing is not 100% accurate. First off, even assuming that all reagents and equipment were effective, and all procedures were followed correctly, a sample could still produce a false negative. How? A negative result simply means that the (potential) viral RNA in the sample did not meet the limit of detection of the test. This could happen early on in the infection (the virus has not significantly replicated yet), or toward the end (the virus has almost been eliminated by the immune system), when the total viral count is lower. There could also be a plethora of problems throughout the testing procedure: the sample could be taken or stored incorrectly, the test kit reagents could be inconsistent, laboratory error, etc. We do not have enough confirmed data to be certain of the accuracy rate of these genetic tests, but apparently similar tests for the flu have only been 50-70% accurate in the past.

Additionally, because genetic testing has not been very widespread in the United States, and the vast majority of people have not been tested, we don’t have the best idea of how many people have actually been infected with the coronavirus. Ideally, this virus has been significantly more widespread than we realized, and we just didn’t know it because people didn’t have bad enough symptoms (or no symptoms at all) to warrant being tested with our limited resources. Also, we can’t use the genetic test to back-test people who have already recovered from the disease, because it only works on people who have an active infection (coronavirus currently in their bodies). This is where another type of test could come in handy: serology testing (or antibody testing).

Serology tests screen for the presence of antibodies, or immunoglobulins, to a particular pathogen. The presence of antibodies indicates exposure to the pathogen, because they are produced to neutralize pathogens and their virulence factors in a variety of ways. Two important classes of immunoglobulins are IgM and IgG. IgM is the first antibody class produced during the first exposure to the pathogen. IgG is the most common antibody present in blood and tissue fluids, but it kicks in a little later than IgM to a new threat. However, it also provides the longest-lasting protection due to its longer half-life. IgM levels will fall as IgG increases, and over time, only IgG antibodies will still be present.

The presence of IgM antibodies suggests a current or very recent infection, because IgM has a shorter half-life (and therefore does not last as long). The presence of IgG antibodies suggests either a long-term infection or a past infection. If someone only has IgM antibodies to a pathogen, then they have been recently infected, and even if they show no symptoms, they are likely infectious. If both IgM and IgG are detected, then that person is probably on the back end of their infection, or has already eliminated it. If someone only has IgG antibodies present, then they have probably recovered from the disease and are no longer infectious. IgG sticks around for a long time and is central to the secondary immune response — aka when you get infected again with the same pathogen — that prevents you from getting sick a second time. If we could test people for COVID-19 specific antibodies, and determine who has been infected and recovered already (IgG positive only), we could selectively send people back out into the workforce and help get the economy going earlier. It could also inform us of asymptomatic carriers (IgM or possibly IgM and IgG positive) who need to be isolated for a little longer, to reduce the spread of disease. Although the presence of antibodies does not necessarily mean that someone is immune, it is a better indicator than none at all, and is a good rule of thumb.

New serology testing is a good development, but we’re still not sure how accurate it is. The new tests coming out are not officially approved by the FDA, and have to put disclaimers on their products. This was done to help speed up the rollout of serological testing, much like skipping regulatory steps in the clinical testing of new treatments. Under normal circumstances, the regulatory process for new tests and treatments would delay public use by many months, if not years. This is not an option, even though there is increased risk with deregulating. The disclaimers essentially say that the tests are not FDA approved, should not be used as a sole indicator of COVID-19 immunity/infection, and may produce a false negative in response to antibodies for a very similar virus. With all of this uncertainty surrounding test accuracy, these new tests don’t seem very encouraging. What is the point of these tests if they are apparently so untrustworthy?

I think that some of these tests will be more on-target than some nay-sayers predict. Especially since there are so many, produced by different companies, there are bound to be some that are better than others, and we will determine which those are. Additionally, if it makes people feel better to know the government bureaucracy is involved, the CDC is also developing its own test to use soon. And ultimately, what are our options if we don’t use widespread testing to (1) better assess the case-fatality and symptomatic case rates, (2) improve our predictive models, and most importantly, (3) assess who is lowest risk for contagious infection (i.e. those who have been infected and recovered)? Either we keep everyone in lockdown, with no clear end (the public needs a game plan with an ultimate deadline to reopen the economy — and we can’t keep adding trillions to our debt to keep the financial demons at bay forever), or we let everyone loose without any decent idea of the unnoticed cases of COVID-19 (while maybe initially keeping the “high risk” under containment). Or some combination thereof. Even if the serology testing is only 50% accurate, that’s 50% more accurate than letting everyone loose without testing. And in combination with molecular testing, we could be even more precise. It will be interesting to see how these tests are incorporated into public policy decisions soon.