"The entire ‘second wave’ is supported solely on the back of a flawed mass PCR [Polymerase Chain Reaction] test, which at industrialized scale was never, in my view and the views of others skilled in PCR, capable of delivering trustworthy results."

- Michael Yeadon, 25 December 2020 (see our Covid Sceptic page)

“For months the lockdowners have been ridiculing sceptics for criticising the PCR test and putting 'cases' in inverted commas. That’s the basis of the smear that we’re covid deniers. Today the @WHO finally acknowledged there’s a problem with the PCR test.”

- Toby Young, 20 January 2021 (see our Covid Sceptic page).

1. The tests have several features that make their results trustworthy. Polymerase Chain Reaction (PCR) tests work by looking for very small amounts of the virus's genetic material (RNA), rather than looking for the virus itself. The accounts are so small that they have to be "amplified" by a chemical chain reaction. Each time a PCR “cycle” is run, the tiny amount of RNA is amplified, and after many cycles have occurred, there should be enough of the RNA present to be detectable. This would give a positive result. Alternatively, if no such RNA is present, the PCR test would come back negative after this amplification process.

The tests usually perform these cycles of amplifying reactions 40-45 times. This has led to concerns among some Covid Sceptics that the tests are amplifying irrelevant—or even non-existent—scraps of RNA. The claim is essentially that so many cycles are run that the test is bound to come up with a positive result eventually - even if no relevant viral RNA is really there. In other words, because of the number of cycles, the test will return far too many false-positive results (we respond to a related argument here).

But the makers and users of the tests have considered this, and built mechanisms into the test to reduce the chance of it coming up with false positives. There are a number of PCR methods currently used in the UK, with most involving "assays" that detect two or more of the virus's genes. Checking whether one gene is present is okay, but if two or more genes have to be present, it substantially reduces the chance of making a mistake in identifying the RNA. The most commonly-used test in the UK looks for three separate genes - in effect triple-checking the result. There are also carefully designed positive and negative controls incorporated into each reaction. Ambiguous cases found at higher cycles—where the chance of a false-positive is higher—are reported as inconclusive, and people are asked to attend for a repeat swab and test.

Not only that, it's simply not the case that most cases are found at higher cycle numbers. Although the cycles are run 45 times, positive results often comes from a lower number of cycles, typically below 30. Contrary to the Sceptics' claims, it is possible to see the number of cycles used to determine positives by UK laboratories. It is far below 45, and drops lower during outbreaks, when a lot more people are infectious and coronavirus RNA fragments are easier to find.

The graph below shows the number of cycles at different times during the pandemic. The vast majority of declared cases (the black dots, which make up 82% of the cases in this sample) were detected on the basis of three genes. As noted above, this makes it much more likely that the coronavirus was in fact present. Also, these cases tend to be found below 30 PCR cycles, once again underlining that the number of cycles is not contributing to a surge of a false-positive cases.

Source: this preprint.

2. There's corroborating evidence. Because the PCR test looks for and magnifies small quantities of genetic material, Covid Sceptics are correct that at higher cycles, for some individuals, a positive PCR test may not indicate that a person is actively infected - some people might just be carrying the dead virus. However, there are strong reasons to think this is not a concern for our understanding of the Covid pandemic. Two such reasons come from corroborating evidence.

First, we can look at clinical studies that use an entirely different method to check for the presence of the virus. One such study took samples from participants and tested them with both PCR and a virus-culturing method, where they checked whether infectious coronavirus could be grown from the sample.

They found that the chance of successfully culturing the virus did decrease with a higher number of PCR cycles (as would be expected), but that anywhere under approximately 27 cycles, around 65-95% of patients were carrying an infectious virus. That is, the vast majority of patients at or below 27 PCR cycles were very likely to have been true (not false) positives for the virus.

This is illustrated by the graph below, which shows the relation between the PCR cycle number (on the x-axis) and the ability to recover an infectious virus from a sample (on the y-axis).

Source: this paper.

Second, people who have fought off the virus tend to quickly flush the genetic material from their system: 14 days after symptom onset, PCR levels will mostly be above 30 and will not generally record a positive result. During the Summer in the UK, soon after the initial Spring wave, at one time 0.08% of Office for National Statistics PCR tests were positive. This was despite around 7% of the population having had the disease just a few months before. If remnant, dead RNA was frequently being picked up to a great extent by PCR tests, we would have seen far higher positivity rates during Summer. We did not.