PCR and the science of the fatality rates of Covid-19
The last blog defined what we mean by infection fatality rate and its importance in determining the seriousness of a disease. Last time, I said I would not get into too many details. For those of you who love numbers and are looking for more details of the science, this blog is for you!
The infection fatality rate is simply the number of deaths it has caused, divided by the number of people who have been infected. First, RT-PCR tests are being used to determine how many people were infected. We’ll look at some of the difficulties in the use of this technology. Second, we need to know how deadly it is among people who were infected with this virus. One might think it is as simple as counting all of the death certificates with Cause of Death as Covid. However, this is complicated by the fact that most of these patients are elderly with all kinds of co-morbidities, so determining what caused their death is not so straight forward.
Infection Fatality Rates
After all of the disclaimers and issues described in my last blog, [LINK] , it is time to get into more nitty gritty science. First, some numbers for the most severe cases. In a study conducted across France last year of over 100,000 hospitalized patients, mortality for COVID-19 was almost three times higher than for Seasonal Flu.
When WHO first declared the pandemic, the reported COVID case fatality rate (the number of known cases divided by the number of deaths) was 6.4%. Since that time, we have been able to be more accurate, once wider testing found many milder cases. Currently, the estimate has dropped closer to 1%.
By summer 2020, we could also start determining infection fatality rate, which is generally several fold lower, rather than just the case fatality rate. The case fatality rate can be problematic in trying to compare one region to another, because it is dependent on all kinds of local variables, not just the severity of illness.
OK, it is time for the big reveal — you’ve been waiting patiently since the last article for the big number. Here it is! On the WHO website in Oct 2020, John Ioannidis (Professor of Epidemiology at Stanford) published that he had calculated the Covid-19 infection fatality rate to be 0.23%. He calculated this from his meta analysis of 60 research studies done around the world. (Other estimates have been somewhat higher : 0.50% — 0.68%.)
As could be expected, fatality rates are often somewhat higher among the elderly. However, in the case of Covid, it is drastically higher.
Ioannidis says that this number drops to 0.05% for healthy people under 70 years of age. Researchers have decided that it makes sense to calculate the different age groups separately, because of the great difference between the elderly and everyone else. As a comparison, the infection fatality rate for Seasonal Flu is generally around 0.02% to 0.06%.
Science Behind Sampling Errors and Inappropriate Testing
This is the section where we’ll delve a little deeper into the issue of how deadly the Covid virus is, and the background of the science to be able to tell. I felt that there are a few sidenotes I must make to clarify those statements I just made, which may be shocking to some readers. Please see my next blog for how others are writing about all of this Covid science-y stuff. So, here are a few things you may not have thought of before, which might give you a clearer understanding of what ‘following the science’ means when it comes to Covid-19.
Testing random people in the street with a low viral load, and low probability of being infected at all means that the test is now being used under radically different conditions from those that the method was validated under. This means you have to re-examine how many false positive results you can expect to get from the test.
The WHO Diagnostic Flowchart specifically says that the PCR test is to be done if the person meets the clinical criteria for Covid-19. NOT everybody!
“Most PCR assays are indicated as an aid for diagnosis, therefore, health care providers must consider any result in combination with timing of sampling, specimen type, assay specifics, clinical observations, patient history, confirmed status of any contacts, and epidemiological information.” — WHO Sept 2020 Report
Note that they tell us these tests are supposed to be as an adjunct to examination of a patient, not instead of.
Because if you are testing a bunch of people who probably don’t have it, this is what happens …
“WHO reminds IVD (in vitro diagnostics like RT-PCR) users that disease prevalence alters the predictive value of test results; as disease prevalence decreases, the risk of false positive increases. This means that the probability that a person who has a positive result (SARS-CoV-2 detected) is truly infected with SARS-CoV-2 (Covid-19) decreases as prevalence decreases, irrespective of the claimed specificity.” — WHO Publications
Note that they tell us that if these tests are done on mostly healthy subjects, their inaccuracy goes way up. Great, right?!
Interpreting a Positive Covid PCR Test
And how are we to interpret a positive Covid PCR test? Here is the answer from the WHO website Jan 13, 2021 update :
“WHO guidance Diagnostic testing for SARS-CoV-2 states that careful interpretation of weak positive results is needed. The cycle threshold (Ct) needed to detect virus is inversely proportional to the patient’s viral load.”
“Careful interpretation of weak positive NAAT (like RT-PCR) results is needed, as some of the assays have shown to produce false signals at high Ct values.” — WHO Publications
So, if there is not much virus in the person, a high number of cycles is needed, leading to false positive test results. Unfortunately, they are somewhat vague on what ‘high’ is, but 40 seems like the absolute maximum. It is equally difficult to find out what testing facilities are using as their threshold number of cycles to be able to find a virus fragment and declare a positive result.
Alright, so this is how to find if a person has virus in them. Is it alive? Will it make them sick? Are they infectious?
Wanted: Dead or Alive
Of course, growing viral cultures from the sample is a sure sign that the person had living virus in them. If it can keep growing a colony, it’s still alive! This takes several days for it to grow, and so this technique can’t be used large scale.
A study conducted in France (corroborated in Korea and Taiwan) comparing the positive RT- PCR test analyses of nasopharyngeal samples with cultured viruses from the subject’s samples found the following results:
- At Ct=25 70% of them showed positive cultures
- At Ct=30 20% of them showed positive cultures
- At Ct=35 3% of them showed positive cultures
A positive result only at a high Ct would indicate a very low viral load. And even if we do know how much viable virus a person has in their body, it’s still unclear how this relates to how infectious they are. There is no test for infectiousness at this point.
A study published in Dec 2020 found that about half of the people who initially test positive for Covid while having no symptoms, eventually get symptoms later. This implies that the other half never get sick.
That’s fine for them, but what about the rest of us? Can we get sick from them?
There has been a lot of speculation about people not having symptoms and spreading the disease. Usually with corona viruses, people are infectious for 2 days before getting symptoms. This is for patients who are pre-symptomatic. A city-wide prevalence study of almost 10 million people in Wuhan found no evidence of asymptomatic transmission.
Since people testing positive are not being followed by a doctor, when someone tests positive on one test with no clinical exam, we have no way of knowing if the patient is asymptomatic throughout the course of infection, or just pre-symptomatic, or has sub-clinical signs at the time of testing, or are possibly post-infection. And, yes, people will still test positive for some time after they are better.
No study has been able to culture a virus from a patient after the 9th day of illness, even though we know they have a high viral load in quantitative PCR tests (ie. the person tests positive).
So, viral load does not seem to be indicative of transmissible live virus.
This suggests that people who don’t get sick are at no risk of infecting anyone, but I’d like to see corroborating research studies, just to be sure. Because that’s what scientists do!
Wouldn’t it be nice if there was a quick test that would tell us that everyone who is let in to the stadium wasn’t infectious? There has been a lot of media pressure for governments to use such quick tests.
If a Rapid Antigen test shows a positive result, these tests are reliable to say that the patient is indeed positive.
Unfortunately, these quick tests picked up only about 50% of those who tested positive in PCR. So assuming the PCR test is accurate, the quick test is going to let in twice the number of infected people that the PCR test would. This all of a sudden makes it much less interesting.
Then there are the logistics of testing everyone who is going to the airport, cinema, football game. To solve this problem, some are advocating that pharmacy employees do the sampling for Covid tests.
Perhaps it is not surprising to hear that in a scientific study done in Norway, the newly trained employees were found to catch 20% fewer positive results than experienced nurses or research technicians. This is not an issue of a weakness in the technology, just a matter of technique and the benefit of experienced hands.
Of course, this virus is a serious threat. And the state of the science changes over time. But if some people say it is ten times worse than the seasonal flu, ask them which demographic they are talking about. In the thousands of hours of Covid coverage you have heard in the past year, you have probably not heard about the major impact these scientific details have on the story. For real keeners, you can go back and check out all of my links for even more details.
I haven’t seen the media report the Ct of the PCR test when they give case numbers, or even which test was used. And I’ve looked! When you use one test in August and another one in September, you can’t just compare the results from each, one to one. If a test that a scientist came up with for one purpose, is being used for a different purpose, it may result in false positives, and not fit for purpose.
My point in telling you all this, is not so you become better chemists. That probably is not going to make your life any better. BUT, it is becoming more and more important to be scientifically literate.
Follow me to see the next Blog: What the Media tells us about Covid — How to understand the messaging of Covid-19.