Are we headed to dangerous levels of overdiagnosis by interpreting a test in a way that labels people as sick and infectious when they may be neither?

EIGHT MONTHS INTO THE PANDEMIC, here are some BC numbers to think about:

5,071,000: Population of BC (est. 2019)

38,471: Typical number of BC deaths in a single year (2019)

132: Number of BC deaths, on average, everyday. (est. 2019)

274: Number of days between Jan. 15 and Oct. 14, 2020

36,168: Estimated number of total deaths in BC between Jan 15 and Oct 14, 2020

250: Number of deaths in BC attributed to COVID-19 up to Oct 14, 2020

0.69: Percentage of total BC deaths over 8 months possibly due to COVID

10,836: Number of “laboratory confirmed” cases of COVID up to Oct 16, 2020

691,741: Number of SARS-CoV-2 tests Jan. 15– Oct 13, 2020

1.82%: Proportion of COVID-19 tests in BC showing as “positive.”

As COVID’s daily data dump lands on our heads, shaped by scorekeeping, commentary and predictions, it’s pretty easy to get lost in the numbers and what to make of a nasty pathogen circulating in our communities.

What stands out from these numbers?

An extremely low likelihood of death by COVID-19 in BC. Certainly lower than any annual toll of the flu. Certainly lower than the numbers of people who have died from cancers, heart attacks, overdoses, suicides and the myriad of other things that take life every single day. If you take 2019 as an average, 132 people per day die in BC, from all causes. That was the last full year without a pandemic virus.

With less than one person per day dying of COVID in BC, one is tempted to ask if we’re making a mountain out of a molehill. I’m increasingly surprised by the general subservience of the populace and the absence of thoughtful dissent against emergency measures that are undoubtedly causing all kinds of other suffering, wreaking long-term havoc on our society, our livelihoods and our economy.

People are quick to point at our numbers and say what a wonderful job BC public health people are doing, keeping COVID cases down and deaths by COVID at a minimum. We are an obedient lot and so listen to Dr Bonnie, among others, who reminds us to limit contact, wear masks, and control the virus by widespread testing, even if some people have reported how hard the tests are to get. 

Yet, if it is true that the SARS-CoV-2 causes the respiratory disease COVID-19, how much effort has been put into ensuring the virus test is done properly, evaluated thoroughly and adequately interpreted? Basically, can we trust the test?

What is it about the test?

Kary Mullis won the Nobel Prize in chemistry 1993 for inventing the PCR (Polymerase Chain Reaction) test, the test that is now being used to ascertain whether or not a person has COVID-19. His test eventually became the standard test that drew the definitive link between the HIV virus and AIDS.

Ironically, Mullis himself was at the forefront arguing that PCR should not be used as a tool to diagnose the disease. Why? Because even if it could identify the presence of a virus, that detection did not mean the virus was capable of infecting other cells. An eccentric and vocal iconoclast, with a penchant for dropping acid, Mullis went to his grave last year continuing to decry his test being misused to diagnose HIV.

Like many jurisdictions in the world, BC employs RT-PCR to test for COVID. It uses an enzyme called reverse transcriptase to take a piece of RNA (ribonucleic acid) which comes from a swab deep inside the patient’s nose. Adding viral enzymes to the RNA converts it into DNA through what is called Polymerase Chain Reaction. The DNA is turned into billions of copies and a fluorescent signal is added, which, after being run through numerous cycles of heating and cooling, can be detected. This amplification allows the needle in the haystack to be seen.

Here’s where things get interesting: The “Ct” or cycle threshold is the number of cycles needed to see the fluorescent signal. So how many cycles of heating and cooling do you need to determine a definitive “positive” or “negative” result? If you don’t detect the virus after a few dozen cycles does that mean the patient is negative? What if you do more than 30 which many molecular biologists say is like trying to squeeze blood from a stone? There may be detectable virus in that highly cycled sample but it is so small and so dead it’ll never be able to infect others.

I put some questions about BC’s Ct cutoff to a spokesperson from the BC-CDC and here’s what she wrote back: “The cycle threshold number used to diagnose COVID-19 may vary based on the test used but we typically use a cutoff of 35 cycles.” She added that other targets (the RDRP and E gene) and certain assays “use cutoffs of 40 or even more cycles.”

I’m no expert, but I wondered: Shouldn’t they have a constant Ct—because changing it can dramatically change the number of positives? It also makes me wonder that if BC uses a Ct of 35 and Ontario (whose Ct, I’m told, is set at 38) then can this alone explain why BC has a much lower level of positive cases? If some countries set the Ct at 20 (very low) and others set it at 40 (absurdly high), how can one even compare levels of positivity between jurisdictions? This really matters.

I consulted a molecular biologist (who asked me to withhold her name as she works as a provincial government biologist) who said that we have to be very cautious in interpreting these tests because the reverse transcriptase enzyme has poor efficiency in converting RNA to DNA. She told me that if we do over 30 to 35 cycles “we can’t culture a live virus from the sample.” Basically, she added, “a high cycle threshold means we’re finding meaningless fragments that say nothing about the infectivity of the patient.”

This is an expert who uses the RT-PCR test everyday in her work doing forensic science, so I trust she knows its limitations. She was quite forthright in saying that possibly as many as 90 percent of those testing positive for COVID-19 are probably not infectious. Which is to say they may have had “fragments” of the virus, but they couldn’t possibly spread the virus to anyone else.

Is a “positive” test really positive?

Where this is heading is a dangerous level of overdiagnosis. Other commentators have said the rate of false positives might be 50-80 percent.

More testing and more false positives would help explain why deaths and hospitalizations aren’t rising on the same trajectory. It’s because some new “cases” are unlikely infectious or indicative of ill health. 

Again a “positive test” is about declaring a person “infected and infectious” but what follows from that? If we were to say maybe half of those 11,000 people in BC testing “positive”—and therefore have been subject to quarantines, social isolation and stigma—then that’s an awful lot of people who have been unfairly labelled and isolated with a disease they couldn’t possibly transfer to others.

Maybe this comes down to a question of what we value. Is it better to have a non-sick person incorrectly labelled as sick (a false positive) than to have a sick person labelled as not sick (a false negative)?

I can understand the BC CDC’s position, because it’s based on the greatest fear of all, the boogeyman of underdiagnosis. The CDC spokesperson explained it to me this way: “setting the detection threshold too low seems appealing until one misses that early case that can transmit infections to multiple people.” The implication here: we can’t be too careful.

However, with screening people for disease you will always have overdiagnosis and underdiagnosis, and careful testing tries to eliminate the possibilities of both false positives and false negatives. It would seem to me that we are likely doing a serious disservice—to society and our economy—by interpreting a test in a way that labels people as sick and infectious when they may be neither.

Alan Cassels is a drug policy researcher and author in Victoria. He is the author of Seeking Sickness: Medical Screening and the Misguided Hunt for Disease.