Summary
Without PCR testing data mortality and morbidity would not be attributable to the novel virus and if this attribution is false there must therefore be other explanations for the ‘pandemic’.
Whilst scandals about PCR testing are well known a materially important aspect of PCR testing has been given scant attention and that is cross reactivity (or crosstalk). This is where other viruses, such as common colds or flus etc., might trigger a false positive PCR result for SARS-CoV-2.
We analysed the very few studies that used blind samples of colds or flu viruses to undertake ‘mystery shopper’ testing of laboratories using PCR.
In these studies, we found strong evidence for cross reactivity between other competing viruses and the PCR tests sampled, rising to up to 25% for other circulating viral pathogens.
We identified many systematic weaknesses in these studies, finding that numerous implicit tricks had been used to supress evidence of cross reactivity, thus misleading the public into believing PCR tests are reliable.
The only circumstance in which a positive PCR test for SARS-CoV-2 might be considered reasonably diagnostic of infection is where there is an unfeasibly strong prior assumption that competing viruses, such as influenzas and competing coronaviruses, had virtually vanished. As well as acting as a strong and unsupported prior disposition this assumption is circular and is based on test results collected using the same faulty PCR technology.
We conclude that, as the argument there was a genuine pandemic caused by the novel SARS-CoV-2 viral pathogen has largely relied on data from highly inaccurate and unreliable PCR testing, this line of argument is no longer tenable.
Background
Arguments that there was a genuine pandemic caused by the novel SARS-CoV-2 viral pathogen have largely relied on data from PCR testing used for diagnosis and transmission surveillance. Without PCR testing data, published official statistics about infection, hospitalisation, and death rates would not be reliably attributable to the novel virus. However, if this attribution is doubtful, partial or false we need to look to other explanations for the pandemic.
There has been a recent resurgence in interest in the role of PCR testing in establishing whether or not a pandemic can truly be said to have occurred. A prominent expert in PCR testing Dr Kevin McKernan, has argued that PCR testing validated people’s covid morbidity experiences (they got ill) and given this they would never be dissuaded from the belief there was a pandemic – and it was counterproductive to try to do so. Likewise, the online mouse character JikkyLeaks has stated that the PCR tests were ‘not all fake‘ (but makes the unfortunate conflation that any belief that they are fake must come with a belief that ‘viruses don’t exist’, despite these not being logically conjoint propositions).
Readers will know that this substack has previously discussed PCR testing for SARS-CoV-2 at some length. Much of this discussion has been statistical in nature, but we have also uncovered scandals about ‘asymptomatic covid’ and the use of single gene testing. These articles and others relating to PCR testing can all be found here.
However, apart from one article about the smoking man emails we have not dug too deeply into the triggers and causes of false positives and asked: how might these false positives occur? Discussing this issue might appear like we are trawling over old coals, but it is crucial in addressing the question as to whether there was indeed a novel virus, whether PCR testing was honestly detecting it and whether there was a genuine pandemic.
This article focuses on one central aspect of PCR testing which has been given scant attention: cross reactivity.
Cross reactivity or (crosstalk) occurs when other viruses, such as those causing common colds, flu or flu-like illnesses etc. might trigger a positive PCR result for SARS-CoV-2. Other considerations are important, such as operational reliability and cross contamination within labs, the lack of standardised testing etc, but here we will focus solely on cross-reactivity (which, of course, will be synergistic with the foregoing).
The extent to which cross reactivity was occurring is fundamental to the performance of the PCR test, but no one really talks about it. Instead, much discussion has focused solely on false positives, as if there is some uniformity to what we mean by a false test result.
The notion of cross reactivity is fundamental to the performance of the PCR test, as is identifying the ways in which PCR might fail. We can segregate many categories of false positive but the two categories we are going to examine here are:
False positives in the absence of competing pathogens: This is where the SARS-CoV-2 virus is known to be absent, but the test result is positive indicating its presence, despite the fact all other competing viruses or bacterial pathogens are also known to be absent in the sample. Typically, the substance being tested might be water, serum, or human lysate.
False positives in the presence of competing pathogens: This is where the SARS-CoV-2 virus is known to be absent from the sample, but the test result is positive for its presence, but where one or more competing viruses or bacterial pathogens are known to be present in the sample. Normally only one pathogen is presumed present, and the pathogens are normally competing causes of ILI (influenza like illnesses), such as other coronaviruses.
Here we are interested in observational experiments where laboratory tests are conducted using blind samples in idealised conditions, where the substance being tested is supposedly not known to the laboratory undertaking the testing (aka ‘mystery shopper’ tests).
We can calculate the false positive rate (FPR) as:
We will report here on false positive rates in the absence of competing pathogens, FPR-Absence and those in the presence of competing pathogens (the cross reactivity false positive rate), FPR-Presence1.
At this point it is worth reminding readers that in relation to the discussion which follows about “false positive rates” (FPRs), quite small apparent FPRs (say 5%, which might mistakenly be regarded as acceptable) during a period when viral prevalence is low (say, 1%) are not simply a minor aberration. For statistical reasons, as we have explained elsewhere, such a scenario would mean that in practice nearly all the positive results obtained would be false positives.
What studies have been done on cross reactivity?
The notorious Eurosurveillance paper by Corman and Drosten et al, published in January 2020 tested their PCR test to determine whether it could detect the ‘2019 novel coronavirus’ and also assessed its specificity by measuring the rate of cross reaction when exposed to other coronaviruses, including HCoVs 229E, NL63, OC43, HKU1 and MERS-CoV. Identification of the novel 2019 novel coronavirus was based on testing for the E and RdRp gene assays.
Their results were:
“Using the E and RdRp gene assays, we tested a total of 297 clinical samples from patients with respiratory disease from the biobanks of five laboratories that provide diagnostic services (one in Germany, two in the Netherlands, one in Hong Kong, one in the UK). We selected 198 samples from three university medical centres where patients from general and intensive care wards as well as mainly paediatric outpatient departments are seen (Germany, the Netherlands, Hong Kong). The remaining samples were contributed by national public health services performing surveillance studies (RIVM, PHE), with samples mainly submitted by practitioners. The samples contained the broadest range of respiratory agents possible and reflected the general spectrum of virus concentrations encountered in diagnostic laboratories in these countries (Table 2). In total, this testing yielded no false positive outcomes. In four individual test reactions, weak initial reactivity was seen but they were negative upon retesting with the same assay.”
Note that they claim 100% accuracy and zero cross reactivity.
It is worth pointing out that that in four individual test reactions “weak initial reactivity was seen” but that these were re-tested and found to be negative. They revealed their contorted reasoning by stating that:
“These signals were not associated with any particular virus, and for each virus with which initial positive reactivity occurred, there were other samples that contained the same virus at a higher concentration but did not test positive.”
Just retesting the samples which showed initial reactivity is a scientifically and statistically illiterate approach, constituting extreme selection bias; they have not ruled out whether some samples, which showed no initial reactivity, would have shown some cross reactivity had they been retested them as well as the others. So, in practical terms this means there were four false positives.
In April/May 2020 Drosten was involved in another validation trial reported in Matheeussen et al involving hundreds of participating laboratories worldwide and hundreds of tests. Rather than 100% accuracy they reported the following cross reactivity rates:
FPR-Absence – 2.7%
FPR-Presence (HCoV-NL63) – 3.1%
FPR-Presence (HCoV-OC43) – 2.9%
Compared to the Eurosurveillance paper this paper may be considered to reflect more realistic rates of cross reaction.
The smoking man emails revealed statistics from the Instand report suggesting a false positive (cross reactive) rate of 9.34%2, given the primers used (E gene, N gene and ORF1ab), using samples of HCoV-OC43, HCoV-229E and human lysate tested in laboratories worldwide. A statistical analysis of the cross-reactivity rates, assuming any two from three genes must be positive for a PCR positive, and assuming statistical independence, yields numbers significantly less than 9.34%:
FPR-Absence – 0.07%
FPR-Presence (HCoV-229E) – 2.9 %
FPR-Presence (HCoV-OC43) – 0.3%
Note that in both ‘Drosten studies’ the cycle thresholds used to detect competing pathogens were not revealed.
In the UK, single gene testing was used extensively to generate high positivity rates, as reported here. If we consider the false positive rates for the Instand report and ‘relax’ the overall PCR positive criteria to one gene from three (E gene, N gene and ORF1ab), instead of two from three (in breach of WHO guidelines), we get significantly higher results for cross reactivity (again assuming the gene variable are statistically independent):
FPR-Absence – 3.4%
FPR-Presence (HCoV-229E) – 22.9 %
FPR-Presence (HCoV-OC43) – 9.7%
Hence, based on this data we might be able to safely assume that there is evidence to support the contention there was a wave of false positive covid tests in the UK in winter 2020/21, likely partly caused by common colds (and in the UK these should have been officially recognised as false positives).
Australian PCR cross reactivity study
In correspondence from 2023 Dr Kevin McKernan discussed the smoking man emails (one name redacted):
The report he is referring to is written by the Doherty institute in Australia by Tran et al. They tested the use of the Beijing Genomics Institute (BGI) PCR test at three laboratories in June 2020 looking to determine cross reactivity and found that the first laboratory was 100% specific with zero cross reactions found (no surprises here - it was the Doherty institute laboratory). The two other laboratories had specificities reported as 99.1% and 97.5% respectively (and a cross reactivity of 0.9% and 2.5% respectively):
These numbers look reassuring small until you consider that of the 115 ‘negative’ samples only 24 contained competing viruses: HCoV-OC43, RSV, Rhinovirus, Parainfluenza etc. These were presented for test mixed with either human saliva or VTM (virus transport media). The other 91 samples were inactivated cell-culture-grown SARS-CoV-2 (not water, serum or human lysate).
Notice that the total number of samples was 115 across the three laboratories, so to compute FPR-Presence we need to use a denominator of 24 and for FPR-Absence we should use 91. The choice of denominator is therefore crucial in reporting the false positive rate and by including many non-pathogenic samples in the test suite the combined false positive rate can be suppressed. Hence, when we consider the FPR-Presence rate from laboratory 2 and divide by 24, rather than 115, we get FPR-Presence rate of 8%. For laboratory 3 the FPR-Presence rate is the same – 2 false positives from 24 samples also resulting in 8%. These are much higher values than that reported.
Furthermore, four samples from laboratory three were discounted because they breached cycle threshold cutoffs, that only applied to these tests. Here is the extract from the report describing this:
“All samples were negative for SARS-CoV-2 by Laboratory 1. For Laboratory 2, sample 253 (influenza virus A in saliva matrix) and sample 260 (coronavirus-229E in saliva matrix), were excluded from analysis after returning invalid results (internal reference target Ct>32 or not detected). The remaining 22 samples were negative for SARS-CoV-2 by Laboratory 2. For Laboratory 3, sample 259 (coronavirus-OC43 in saliva matrix) and sample 260 (coronavirus 229E in saliva matrix) were excluded from analysis due to their internal reference target having a Ct>32. The remaining 22 samples were negative for SARS-CoV-2 by Laboratory 3. Of note were 4 samples from laboratory 3: sample 236 (parainfluenza type-1 in saliva matrix), sample 246 (adenovirus type-5 in saliva matrix, sample 181 (rhinovirus in VTM matrix) and sample 191 (influenza virus A in VTM matrix) that generated a Ct value above 38 for SARS-CoV-2. These were interpreted as not SARS-CoV-2 positive as per manufacturer’s IFU.”
From the test report we have no idea what an internal reference control target is but, taken at face value, it presents itself as a convenient way to re-categorise false positives as true negatives, and thus improve test performance by subterfuge. When these discounted false positives are included the FPR-Presence, for laboratory 2, rises to 6 false positives from 24 samples – a whopping 25%.
This discounting of false positive samples with high CT values is not matched in the sensitivity analysis branch of the study where SARS-CoV-2 positives were found at CT values as high as 44. These were judged to be false negatives for SARS-CoV-2 because they had CT values > 38, but were subject to confirmatory retesting, whereas false positives were not subject to this retesting.
Dr Kevin McKernan is therefore correct to point out that there was clear crosstalk between competing viruses and the SARS-CoV-2 PCR test at high CT levels. This Tran et al study is by far the most detailed and revealing of the fantastically small number of public studies done into false positivity. It reveals systematic bias in the choice of denominators, leading to hugely inflated specificity rates for the PCR test evaluated, shows how cycle thresholds can be manipulated to reject inconvenient results and how confirmatory retesting might be preferentially done to unfairly boost performance.
Dr McKernan also said:
“And Australia is used as the model for why they are perfect as they had a long stretch of 1:10K positivity off season.”
Given that respiratory viruses are seasonal pathogens this statement is very revealing. Our analysis clearly shows that the absence of false positives ‘off season’, in summer, proves little to nothing about cross reactivity, because competing respiratory pathogens will be largely absent and hence overall PCR specificity can be made to ‘look good’, with a false positive rate close to zero, by simply conducting PCR testing on people uninfected by any virus in summer3. Of course, this situation is not equivalent to ‘in season’ testing when competing pathogens are all in circulation.
Limits to diagnosability and the myth of vanishing flu
The effect of the cross reactive false positive rate on any estimation of the prevalence rate for SARS-CoV-2 is highly significant. Let’s take some of the statistics we have discovered thus far and look at the probability that someone is infected with SARS-CoV-2 given a positive PCR test result.
First, for example, we assume SARS-CoV-2 is circulating in 1% of the population and all other competing respiratory viruses are circulating at a combined 10% rate: hence 89% of people are virus free. We can use Bayes Theorem to calculate the chance of someone being infected with SARS-CoV-2 when they have a positive PCR test. Let’s further assume a high sensitivity for the PCR test, 99%, with an FPR-Presence of 25%, and FPR-Absence of 0.07% (so those with a competing ILI will have a 25% chance of showing a false positive and 0.07% of people uninfected, with either, will also show a false positive4). Using Bayes’ the calculated result is only a 28% chance that the person tested actually has SARS-CoV-25, as shown in Bayesian model a).
In this example a SARS-CoV-2 PCR positive test result might only be considered ‘reasonably’ diagnostic in the most favourable circumstances where, if and only if, we assume competing ILIs have almost completely disappeared. For example, if we flip the prevalence of ILIs with SARS-CoV-2 where competing respiratory viruses are circulating at a combined rate of 1% and SARS-CoV-2 is circulating with a prevalence rate of 10% then, again using Bayes’ the calculated result would be a 97% chance that the person tested has SARS-CoV-2, as shown in Bayesian model b).
Model b) shows, in Bayesian terms, that to be near certain in the diagnosis of SARS-CoV-2 infection you would have to hold the prior belief that ILIs had all but vanished and that SARS-CoV-2 was exclusively dominant. However, we have argued elsewhere that there is no reason to believe that, caused by the supposed circulation of SARS-CoV-2, other competing ILIs simply disappeared during 2020 and 2021.
The reported global downward trend in competing bacterial and viral infections during spring 2020 is exemplified by this chart, published by the BioFire syndromic trends reporting system, showing the percentage detection rate of pathogens in the USA from 2019 through to 2021. We have argued here that the proposed hypothesis of viral interference between SARS-CoV-2 and influenza or other viruses to explain why these competing viruses simply disappeared only to reappear a few years later is just not credible, with changes in testing being a much more likely explanation.
Finally, in the absence of an available and deployed gold standard, these arguments are circular because the supposed downward trend in the prevalence of competing ILIs are themselves dependent on the same PCR testing technology. If this technology is deployed in a systematically biased way the results will be misleading. One example of this is the startling revelation, by RemnantMD, that for one popular PCR multiplex test which claimed to be able to differentiate between influenza A/B and SARS-CoV-2, a negative SARS-CoV-2 result means then any accompanying negative tests for influenza should be considered presumptive. This was explained in their documentation as follows:
So, if a sample that is tested is positive for SARS-CoV-2, it is assumed to be negative for Influenza.
The reader should note that there are related important issues that would also suppress the signal from influenza such as the timing and availability of influenza testing, the role of quarantines and confirmation biased baked into the influenzas testing protocol.
Discussion and conclusions
We analysed the very few studies that used blind colds or flu samples to undertake ‘mystery shopper’ testing. In these studies, we found strong evidence for cross reactivity between other competing viruses and the PCR tests sampled, rising to up to 25% for other circulating pathogens.
We identified many systematic weaknesses in these studies, finding that numerous implicit tricks had been used to supress evidence of cross reactivity, thus misleading the public into believing PCR tests are reliable. We found that the only circumstance in which a positive PCR test for SARS-CoV-2 might be considered strongly diagnostic of infection is where there is an unfeasibly strong prior assumption that competing viruses, such as influenzas and coronaviruses, had virtually vanished. As well as acting as a strong and unsupported prior disposition this assumption is circular and is based on test results collected using the same faulty PCR technology.
We therefore conclude that, as the argument there was a genuine pandemic caused by the novel SARS-CoV-2 viral pathogen has largely relied on data from highly inaccurate and unreliable PCR testing, this line of argument is not actually tenable.
There are other issues we have not touched on here relating to genetic sequencing etc. We have also not commented on the fact that the sample sizes used in these studies is shockingly small. If we took this into account and computed confidence intervals and performed confidence tests our results would likely be even worse, leading to an inevitable conclusion that PCR testing is not particularly specific and suffers from irreducible cross reactivity and false positive inaccuracies.
Likewise, it might be tempting to argue that even if PCR testing was a failure, then the available alternative, in the form of lateral flow testing (LFT), clearly showed a pandemic was taking place. But this is a moot point because LFTs were not available during much of 2020, and this initial phase of the pandemic was entirely driven by PCR testing. By the end the end of 2020 when LFTs were being widely used the accuracy gap between PCR and LTF positivity was significant. With PCR at 18% versus 4% positivity for LFTs, in week 53 of 2020 in England, it was clear PCR was over inflating the positivity results (see our article here).
Our evidence presented here about PCR’s ability to identify a novel deadly virus needs to be considered in combination with more recent revelations made about supposed uniqueness of symptoms about ‘spikeopathy’, as described here. A pivotal point we made in that article is just as relevant here: objective confirmation cannot be obtained by the PCR test because of the inability of swabs to reliably collect and identify causative agents (as reported by the CDC EPIC study in two 2015 NEJM articles - one done on adults and one on children). Hence, a positive result gained from a sample taken from the upper throat or the nose does not mean an infection in the lung is caused by the detected pathogen. On the face of it this evidence somewhat renders all arguments about PCR irrelevant.
Dr Kevin McKernan has recently espoused the view that that PCR testing validated people’s covid morbidity experiences. He may be correct about current public opinion, but its relevance is questionable here given that, in 2023, he was scientifically correct when he said there is ‘late crosstalk with other viruses’, and his statement is relevant to all PCR testing undertaken up to and including that year.
Without PCR testing data, published official statistics about infection, hospitalisation, and death rates would not be attributable to the novel virus. This, and many of our other analyses have confirmed this attribution is largely false and there must therefore be other explanations for the ‘pandemic’. Such explanations need not include the conclusion there was ‘no virus’.
Note that this is not the same as the false discovery rate, which is false positives divided by the total of false positives and true positives. The false positive rate is equal to the complement of the specificity of the test.
This was calculated by simply totalling frequencies of false positives for single genes.
In the UK confirmatory testing was only performed was when very few people were ill and few were being tested.
In these examples we have used the 0.07% FPR-Absence from the Instand report, and 25% FPR-Presence as calculated from the Tran et al study. This balances the propensity of testing to generate cross reactive false positives on season and off.
Note that these calculations ignore symptoms and do so because the studies we cite are based solely on samples and not people, hence even with positive tests people might be asymptomatic or post symptomatic or might never develop a reaction to the virus.
On a different, but related subject, I think many of us would also greatly appreciate your unbiased professional opinion on the ONS changes to the methodology for calculating excess deaths? I'm sure you must be looking at this already in view of it's trend changes and so reduction in excess deaths and attendent reduction in percentage increases over the "Vaccine years" as a result of higher figures in the "Pre-vaccine" years? Except 2019 which seems like an unexplainable outlier OR was set to deceive in advance?
Why this?, why now? seem relevant questions, particularly since this change now means that UK data will no longer be comparable with ROW data, which still uses the simple moving average method. That's just an accidental coincidence, I'm sure?
Cross reactivity, if I'm thinking about this correctly, sure could explain why cold and flu and every other illness of its type virtually "disappeared" during the pandemic. Meaning, we still had the same numbers of cold and flu but we're instead blamed on covid -substantially inflating its number and thereby creating its pandemic. Deducing that, means a pandemic never really happened but instead what may have actually occurred was the usual seasonal viruses alongside the addition of covid substantially (and falsely)increasing its numbers of positive PCR.