It will be interesting to read the plethora of books that will be written in the future about the Covid pandemic and our response to it. Hopefully level heads will prevail and we will learn how we royally turned the Philips head clockwise in the canine. One area I believe will be scrutinized is the projection model we used to determine the spread and severity of the virus, which spilled into public policy in the form of the fantastically popular school closures and lockdowns.
We use models to help us understand and predict complex systems so we can respond to scenarios presented by the system. Models are built on assumptions. Since there are many unknown variables in complex systems, you have to start somewhere, so we take educated guesses at the individual variables which are based on the best available data. One of the variables in the projection models for Covid is how many people are testing positive for the disease. Some of you had the “brain-tickler” test that went up both nostrils and felt for all the world like a caterpillar was chiseling his initials into your septum. That was a PCR test.
I want to describe what a PCR test is, how it works, and how some grey areas can potentially affect numbers and therefore models. This one will be a bit off the path for me. There will not be any spiritual yummies; it’s just the data. Mostly because I find it interesting. Secondly, because there are any number of ways in which stats and data points can be smooshed around, and a bit of digging is usually necessary to shed light on what a particular statistic is actually measuring. As you will see in the end, I am not implying there was any monkey business, or monkey shines, or monkey shoe-shining business going on. This is only for posterity.
So first, a little biochem background.
A virus can be DNA based or RNA based. In simple terms, DNA is a double helix strand of genetic data and RNA is a single strand. DNA comes with its very own Grammarly subscription, and has the ability to spell check itself when it replicates, correcting inappropriate nucleotide sequences, thereby limiting potentially lethal (to the virus) mutations and preserving the integrity of its genetic code. RNA doesn’t know what it wants to be in life, and therefore has no spell check mechanism so genetic mistakes are missed, giving the RNA virus a much greater potential to mutate. This can be lethal to the virus, or can result in randomly variation which confers a survival benefit. SARS COV-2 is an RNA based virus and the odious Delta variant is one such example of a beneficial (to the virus) mutation. One hallmark characteristic of viruses, particularly RNA based ones, is that they mutate with abandon and are therefore very difficult to get a lid on.
The other quirky thing about viruses is that, unlike bacteria, they are technically not alive because they cannot replicate on their own; they must parasitize off a host. The process of how this is done is important to understand. The SARS COV-2 virus is an RNA strand surrounded by a protein coat called a capsid. This capsid has proteins sticking out of it making it look all spikey, like a kitten after a bath and a good shake. These proteins have a certain shape and magnetic affinity which finds other proteins on your cell’s surfaces that match these spike proteins in shape and affinity, like a magnetic puzzle. When this happens, a door to your cell is unlocked and your cell membrane fuses with the viruses, which poops its RNA into your cell’s protoplasm. From there the RNA heads to the cell’s nucleus where it parasitizes your hard earned DNA to make more of itself (which makes me think perhaps God gave us the coronavirus as an epic analogy for communism. But I digress). After multiple copies of the virus are made, they move out into the jelly of the cell, reassemble their capsids with spike proteins and leave the cell to seek their fortune.
Okay, so far so good. We know what an RNA virus is and how it replicates. On to the PCR test.
A polymerase chain reaction test (PCR test) is a very common test. We use it in the ER frequently for simple stuff like testing for flu or strep throat. It was invented by Carey Mullis as a research tool, for which he received the Nobel prize in Chemistry in 1993. It has the ability to amplify certain fragments of DNA or RNA so they can be studied easier. Most of the mapping techniques used in the Human Genome Project, which mapped the entire human genome, used the PCR process. So it’s pretty dope, as the kids say.
The purpose of the PCR process is to find certain segments of a genetic code and photocopy them. This makes the code easier to study simply by increasing the sheer number of the segments. Let’s say I’m at an arcade with my daughter and she spies one of those crane games where you need to maneuver a claw and release it, hoping it closes around the toy you want. She sees a cute stuffed tiger pressed against the window, near the bottom, and really wants it to play with. It would be very difficult, and potentially very expensive, to maneuver the crane to just the right spot and fish out the right toy in a disorganized pile of other toys that do not interest her. So since I am a certified genius, I develop a mechanism which I introduce to the game that finds the toy tiger and photocopies it – makes an exact replication. This process of doubling is called a cycle. Now there are two copies of the tiger after one cycle, but this is still scant compared to the total number of toys. So I cycle the mechanism again to double the already doubled number of tigers; now there are four tigers. Then I do this again and again until, after 30 cycles, there are millions of stuffed toy tigers. Now when I use the crane I will be able to pull multiple copies of this one particular toy. The mechanism I developed has aided my ability to choose the specific toy I wanted so my daughter could play with it. This is what the PCR process does. It photocopies selected segments of a genetic code, doubling the number of segments with each cycle, so as to make that segment easier to play with – to study.
But let’s say I want to change my crane/claw test. I have a hundred crane/claw games and I want to know which machines have a toy tiger and which do not. How could I do this? How can I turn this process I have developed into a diagnostic test? Since I am now only interested in the question Is this machine ‘positive’ for a toy tiger?, I could do it by adding to my photocopying mechanism a glow-in-the-dark collar which would attach to each tiger. So now every time I cycle the test, if there is a toy tiger present in the machine, it will be replicated and a glow-in-the-dark collar will be attached onto the toy to make its presence easier to detect. Since I really want to find the machines with the toy tigers, I can cycle my test a bunch of times to make those machines which have the tiger really, really bright. I take with me a handy gadget that can detect glow-in-the-dark light which I can adjust to make it more or less sensitive. Then I can turn off the lights, walk by all the machines, and using my light-detecting gadget, see which machines are emitting enough light to trigger “tiger positive.”
What was originally a process I developed to make it easier to get a tiger for my daughter to play with is now a test to see if there are tigers present in any given crane/claw machine at all. This is an important change. Since we have changed the test to only give us a yes or no answer, certain complications can arise. Is it okay with you if I continue to tax your imagination?
Think of similar crane/claw machines you have seen in real life. There are myriad different types of stuffed critters, and many times there are several toys of the same critter. In the same way, my toy tiger is one of many stuffed critters in the machines – there are dogs, aliens, snakes, monkeys, babies, spiders and, importantly, other tigers. Before I even begin my test, some machines may have a bunch of the toy tigers, some have a few, some have none. If after only a few cycles of the test I reach the glow-in-the-dark threshold where my gadget registers a “positive” result, this means there was a bunch of tigers in the machine to begin with. If the light threshold is not met, the test is negative. If positive, I can use the number of cycles it took to reach that threshold to tell me how many tigers were in the machine in the beginning. So if there were one hundred toy tigers in a machine before I began cycling the test, then it would take fewer cycles for my gadget to detect the glow-in-the-dark collars than if there were only a few. The fewer original tigers, the more cycles it takes to reach the threshold, because it would need more cycles to arrive at that threshold.
Now, for the sake of illustration, let’s pretend I go into an arcade with 100 crane/claw machines and I want to see which ones I could get a tiger out of. So I cycle my test on all the machines, say, 24 times. Some of the machines would be super bright with glow-in-the-dark tigers, some would be less bright and some not at all. I can clearly see which machines are “tiger positive” and even though some are less bright, they still trigger my glow-in-the-dark detector; I could get a tiger out of them.
But while I am in the arcade, my elbow accidentally bumps my testing mechanism and I cycle my test a dozen more times, and… what’s this? A bunch of machines that were dark now are registering as “tiger positive”! What happened? To figure this out, I open up those machines to find that my over cycling of the test had incidentally registered a machine positive because they had only pieces of tiger, like a leg or head, not the whole stuffy. Because I had cycled my test so many times, the accumulated pieces grew to a size where it triggered a glow-in-the-dark positive. The number of cycles seems to be important in determining how easy it is to get a tiger from a machine or even if there are any tigers in there at all.
Let’s tie this into the test for SARS CoV-2.
The actual test used to detect SARS CoV-2 is called a quantitative fluorescence-based reverse transcriptase polymerase chain reaction test (RT-qPCR). Briefly, the test takes the SARS CoV-2 RNA, makes it into a DNA strand, then applies the polymerase reaction which photocopies the surrogate DNA. The test does not check for a whole, intact virus, complete with protein capsid and all that, only a viral segment of RNA. A fluorescent marker is added which attaches to the photocopied segments to make them more noticeable. With each cycle, the light from the fluorescence intensifies, since there is a doubling of the number of copies with each cycle. If fewer cycles are needed to reach the fluorescence threshold that means there was a greater starting amount of viral segments present. Conversely, if it takes more cycles to reach the same threshold, the fewer viral segments it begins with. This number of cycles needed to reach fluorescent intensity is called the cycle threshold or Ct.
But here is a potential issue: there is no standardized Ct by which a test is positive or negative. Each testing lab and test manufacturer has different Ct thresholds. Since there is no standard Ct, different countries and provinces use different cycle thresholds to determine a positive or negative test. In Canada, Newfoundland uses a Ct of 33, and Quebec, 45. Remember that we are doubling the light intensity with each cycle. So going from 25 cycles to 45 is an increase of billions which might not indicate a “live” infectious SARS CoV-2 (Oral Health Group, 2021).
Prior to PCR testing, the previous gold standard was viral culture, which means they squirt some of the virus into a cell and see if it proliferates. PCR has a broader application, is faster and more sensitive, but is not as good at identifying infectivity. Infectivity means what it sounds like: the ability of a virus to cause a disease in another person. To test the ability of PCR to discriminate for infectivity, a study compared PCR tests at different Ct and then subjected those same samples to cell cultures. The results published in Clinical Infectious Disease in November 2020 stated:
These results demonstrate that infectivity (as defined by growth in cell culture) is significantly reduced when RT-PCR Ct values are > 24. For every 1-unit increase in Ct, the odds ratio for infectivity decreased by 32%. (Bullard, Jared, et al, 2020)
What this is saying is that if someone tests positive with a PCR test, for every cycle after the 24th the cell cultures grew less and less virus, meaning they were less infectious. Or to make the point more obvious, an individual who tested positive at Ct 10 would be more infectious, whereas someone who was positive at Ct 40 would likely be non-infectious because their viral count is so low, needing billions of amplification to reach Ct. This study suggests that cycling the mechanism more than 24 times will can make the test hypersensitive, detecting particulates of viral segments which are not live or infectious, but have simply been cycled so many times so as to reach the light intensity threshold and therefore register as positive. Dr. Anthony Fauci stated in an interview in July 2020 that a Ct of over 35 does not reliably indicate a live and infectious SARS CoV-2 infection. The CDC accepts a Ct of 40 (CDC, 2020 p.35). The CDC also released a document which identified several limitations of the RT-qPCR test. Among these limitations is that the presence of viral RNA doesn’t mean it is an infectious virus, it may simply be amplified to an extent which makes it appear an active infection (CDC, 2020 p. 37)
This means that in theory a negative test at Ct 24 could be converted to a positive test at Ct 40, simply by raising the cycle threshold. So using the Ct of different provinces in the same country, a person who tests negative in Newfoundland at Ct 33 would test positive in Quebec with Ct 45. Since the RT-qPCR test does not indicate what the Ct is but only if the test is positive or negative, there is potential for assuming greater infectivity than there really is.
The CDC as well as other research organizations say that the Ct cannot be correlated to how sick or infectious someone is. In other words, a doctor wouldn’t treat a person who was positive at Ct 10 more aggressively that Ct 40 – treatment depends on how sick the person is that is in front of you. But it does appear true based on the study above that positive tests over a certain Ct correspond to a lessening of infectivity in cell cultures.
Another giant study from the UK came to similar conclusions
Ct value was an important determinant of PCR-positive results in contacts, with an approximately linear decline as Ct value increased…(Lee, et al, 2021).
This leads me to believe there is at least a gross association between higher Ct and lessening infectivity. And if that threshold where infectivity begins to drop off precipitously is around Ct 24 and most tests are running significantly higher than that, perhaps the PCR is a variable that needs some fine tuning for future models.
I am not espousing any kind of conspiracy theory. Not at all. The fact that we have this kind of technology is quite incredible and I am thankful for it. Much good has come from it. I am also not saying the PCR test is unreliable per se. But it will be interesting as we look back on this whole episode and see how we could have done things differently, where mistakes were made, what areas conflated to result in improper modeling, and what we can do differently if, God forbid, we have another viral pandemic.
References
Bullard, Jared, et al. “Predicting Infectious Severe Acute Respiratory Syndrome Coronavirus 2 from Diagnostic Samples.” Clinical Infectious Diseases, vol. 71, no. 10, 2020, pp. 2663–2666., doi:10.1093/cid/ciaa638.
CDC 2019-Novel Coronavirus (2019-nC0V) Real-Time RT-PCR Diagnostic Panel. Available at: https://www.fda.gov/media/134922/download
Lee, Lennard Y, et al. “Severe Acute Respiratory Syndrome Coronavirus 2 (Sars-Cov-2) Infectivity by Viral Load, s Gene Variants and Demographic Factors, and the Utility of Lateral Flow Devices to Prevent Transmission.” Clinical Infectious Diseases, 2021, doi:10.1093/cid/ciab421.
“The Problems with the Covid-19 Test: A Necessary Understanding.” Oral Health Group, 8 Feb. 2021, http://www.oralhealthgroup.com/features/the-problems-with-the-covid-19-test-a-necessary-understanding/.
Thought Butter 